<%BANNER%>

Dynamic Project Management Methodology

Permanent Link: http://ufdc.ufl.edu/UFE0022522/00001

Material Information

Title: Dynamic Project Management Methodology Managing Schedule Compression
Physical Description: 1 online resource (141 p.)
Language: english
Creator: Lee, Jaesung
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: automation, compression, constraint, construction, dynamics, lean, management, pressure, schedule, scheduling, system
Civil and Coastal Engineering -- Dissertations, Academic -- UF
Genre: Civil Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Despite aggressive efforts of project managers to maintain a project schedule or to recover from a lapsed schedule, delays and cost overruns have become routine phenomenon at many construction projects. This research proposes a proactive schedule compression method to reduce the expected project completion time by removing latent lazy time caused by fallible scheduling and constraints that impose non-value-added effects on project. An additional objective of this research is to develop a systematic environmental model, the Dynamic Project Management Model under Schedule Compression (DPM), to improve performance against latency and complexities of design and construction projects in schedule compression. Developing and implementing the DPM model will result in the following research outcomes: the association of schedule compression and project components with the construction process and their effects on construction performance; detecting and eliminating latent lazy time that project has potentially; managing latency caused by schedule compression; understanding of schedule compression frameworks; and developing system dynamics-based schedule compression model to improve performance under the condition of latency and complexities of design and construction project.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jaesung Lee.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Ellis, Ralph D.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2009-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022522:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022522/00001

Material Information

Title: Dynamic Project Management Methodology Managing Schedule Compression
Physical Description: 1 online resource (141 p.)
Language: english
Creator: Lee, Jaesung
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: automation, compression, constraint, construction, dynamics, lean, management, pressure, schedule, scheduling, system
Civil and Coastal Engineering -- Dissertations, Academic -- UF
Genre: Civil Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Despite aggressive efforts of project managers to maintain a project schedule or to recover from a lapsed schedule, delays and cost overruns have become routine phenomenon at many construction projects. This research proposes a proactive schedule compression method to reduce the expected project completion time by removing latent lazy time caused by fallible scheduling and constraints that impose non-value-added effects on project. An additional objective of this research is to develop a systematic environmental model, the Dynamic Project Management Model under Schedule Compression (DPM), to improve performance against latency and complexities of design and construction projects in schedule compression. Developing and implementing the DPM model will result in the following research outcomes: the association of schedule compression and project components with the construction process and their effects on construction performance; detecting and eliminating latent lazy time that project has potentially; managing latency caused by schedule compression; understanding of schedule compression frameworks; and developing system dynamics-based schedule compression model to improve performance under the condition of latency and complexities of design and construction project.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jaesung Lee.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Ellis, Ralph D.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2009-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022522:00001


This item has the following downloads:


Full Text

PAGE 1

1 DYNAMIC PROJECT MANAGE MENT METHODOLOGY: MANAGING SCHEDULE COMPRESSION By JAESUNG LEE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Jaesung Lee

PAGE 3

3 To my mother and father

PAGE 4

4 ACKNOWLEDGMENTS What I would like to say is only one: this work has not been done by me My parents, their constant encouragement was a source of inspir ation and their unconditional love and endless support have driven me to never give up. This wo rk is the result of thei r smiles, tears, caring, patience, belief, and sacrifice. Words cannot expr ess my appreciation. What I can say is only that I love them. I love my mother and father. I would like to thank Dr. Ralph D. Ellis for se rving as my advisor. His continued support and availability during my research and studies were what made this possible. I thank Dr. Charles Glagola and Dr. Zohar Herbsman for thei r guidance, advice, and real world perspective on academic problems during my graduate studies, as well as Dr. Ian Flood for his assistance and input during this research. They not only helped with individual advi ces, but also gave me pleasant memories. In addition, I would like to acknow ledge the contribution to this thesis by Lorie A. Wilson, Johathan M. Duazo, and Abel Sierra at Florida Department of Transportation. Also, I would like to thank Dr. SangHyun Lee, his advice and achievement carried the seed of this work and entrusted it to me. I have been blessed to receive guidan ce and support from many whose advice and consolation have reminded me not to stand by myself. These generous souls include Hyuntae Jung, Jaeyun Kim, Yusun Chang, Younghwan Byun, Youngjun Park, Dohyung Ha, Jeawan Hwang, Hoki Moon, my brother Junoo, and their fam ilies. I am very grateful to them. And, I would like to my deep appreciation to my si ster, Sujin Lee, and her husband, Youngjae Lee. Last, I offer my heartfelt thanks to Carol yn Tuttle who was sitting beside me in long moments of absolute solitude.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4LIST OF TABLES................................................................................................................. ..........7LIST OF FIGURES.........................................................................................................................8LIST OF DEFINITIONS...............................................................................................................11ABSTRACT...................................................................................................................................13 CHAPTER 1 INTRODUCTION..................................................................................................................142 RESEARCH APPROACH.....................................................................................................182.1 Research Assumptions..................................................................................................18 2.1.1 The Yerkes-Dodson Law................................................................................18 2.1.2 Parkinsons Law.............................................................................................. 192.2 Research Applications..................................................................................................202.3 Methodology Overview................................................................................................22 2.3.1 Analysis...........................................................................................................23 2.3.2 Development...................................................................................................24 2.3.3 Validation........................................................................................................253 SCHEDULE COMPRESSION MANAGEMENT FRAMEWORK...................................... 263.1 Literature Review........................................................................................................263.2 Feedback Process of Schedule Compression............................................................... 293.3 Latent Lazy Time and Latency....................................................................................303.4 Relationship of Activities............................................................................................32 3.4.1 Evolution and Sensitivity................................................................................ 32 3.4.2 Dependency.....................................................................................................343.5 Schedule Compression Management Framework........................................................ 36 3.5.1 Internal Project Management Framework...................................................... 37 3.5.2 External Project Management Framework..................................................... 444 CONSTRAINTS MANAGEMENT FRAMEWORK............................................................ 464.1 Applied Concepts....................................................................................................... ...47 4.1.2 Theory of Constraints.....................................................................................47 4.1.2 Key Constraints Analysis................................................................................ 48 4.1.3 Shielding Production....................................................................................... 494.2 Constraints Management Framework........................................................................... 50

PAGE 6

6 5 SYSTEM DYNAMICS-BASED PROJECT MANAGEMENT MODEL............................. 535.1 System Dynamics.........................................................................................................535.2 Modeling Process..........................................................................................................575.3 Model Boundary...........................................................................................................595.4 Feedback Processes of Construction under Schedule Compression.............................605.5 System Dynamics-Based Project Management Model.................................................70 5.5.1 Basic Work Rate............................................................................................. 73 5.5.2 Constraints Management Process................................................................... 74 5.5.3 Schedule Compression Management Process................................................. 76 5.5.4 Quality Management and Work Completed................................................... 805.6 Development of Dynamic Construction Project Model............................................... 826 APPLICATIONS................................................................................................................... .846.1 Dynamic Project Management Model Application...................................................... 85 6.1.1 State Route 25.................................................................................................85 6.1.2 Validation........................................................................................................86 6.1.3 Analysis...........................................................................................................886.2 Policy Implications.......................................................................................................937 CONCLUSIONS.................................................................................................................... 997.1 Applicability of Dynamic Project Management Model................................................ 997.2 Future Research..........................................................................................................101 7.2.1 Standardization of Co mpression Concepts and Methods............................. 102 7.2.2 Computerized Dynami c Project Management Model ................................... 104APPENDIX A STRUCTURES OF SYSTEM DYNAMICS MODEL........................................................106B EQUATIONS OF SYSTEM DYNAMICS MODEL........................................................... 117C INPUTS FOR SIMULATION..............................................................................................127LIST OF REFERENCES.............................................................................................................132BIOGRAPHICAL SKETCH.......................................................................................................141

PAGE 7

7 LIST OF TABLES Table page 5-1 Comparison of the traditional appr oach and system dynam ics approach.......................... 555-2 Modeling process........................................................................................................... ....585-3 Model boundary chart........................................................................................................596-1 Example of input variables for simulation......................................................................... 866-2 Policy analysis on SCT reliability......................................................................................987-1 Effectiveness of methods of sc heduling: schedule value rating...................................... 1027-2 Most effective planned and unpla nned schedule compression methodologies............... 103C-1 Input variables for simulation: SR-25.............................................................................. 127C-2 Input variables for simulation: SR-25 II.......................................................................... 130

PAGE 8

8 LIST OF FIGURES Figure page 1-1 Proactive mannered scheduling......................................................................................... 162-1 The Yerkes-Dodson Law...................................................................................................182-2 Parkinsons Law............................................................................................................ .....202-3 Time reduction by elimin ating latent lazy time................................................................. 212-4 Worse performance by schedule pressure.......................................................................... 222-5 Concept of dynamic project management model............................................................... 232-6 Research methodology....................................................................................................... 243-1 Conceptual feedback loop under schedule compression management.............................. 303-2 Effect of latent lazy time and latency................................................................................ 313-3 Information dependency.................................................................................................... 333-4 Determinants of evolution and sensitivity......................................................................... 343-5 Internal dependency........................................................................................................ ...353-6 External dependency........................................................................................................ ..353-7 Internal management framework.......................................................................................383-8 Evolution loop............................................................................................................. .......413-9 External management framework...................................................................................... 433-10 Extended project management framework........................................................................ 444-1 Process of on-going improvement.....................................................................................484-2 Shielding production....................................................................................................... ...504-3 Constraints management framework................................................................................. 514-4 Generation from stabilized environment........................................................................... 525-1 Concise causal loop diagram of schedule compression and LLT...................................... 605-2 Feedback process of general construction......................................................................... 61

PAGE 9

9 5-3 LLT generation loop..........................................................................................................645-4 Internal evolution and sensitivity loop............................................................................... 655-5 Schedule compression management loop.......................................................................... 665-6 Quality management loop..................................................................................................675-7 Causal loop diagram of constructi on process under schedule compression...................... 685-8 Development processes in the model structure.................................................................. 715-9 Initial work rate..................................................................................................................745-10 Constraints management process....................................................................................... 755-11 Work rates relationship by schedule compression............................................................. 785-12 Schedule compression management process.....................................................................795-13 Generic construction process st ructure under schedule compression................................ 815-14 Schema of dynamic pr oject management model............................................................... 836-1 SR-25 project location..................................................................................................... ..856-2 Percentage of work complete of CPM, actual work, and DPM (SR-25)...........................876-3 Work progress rate......................................................................................................... ....886-4 Net redesigned work and net work to be checked.............................................................896-5 Total work awaiting redesign............................................................................................. 896-6 LLT and latency effects on productivity............................................................................ 906-7 Total schedule compression effect on productivity........................................................... 916-8 CPM vs. DPM work rates..................................................................................................926-9 SR-25 II project location.................................................................................................. ..936-10 Percentage of work complete of CPM, actual work, and DPM (SR-25 II).......................946-11 Percentage of work comp lete according to SCT time....................................................... 956-12 Percentage of work complete vs. SCT reliability.............................................................. 966-13 Work rate vs. SCT reliability............................................................................................. 97

PAGE 10

10 6-14 Net redesigned work and net work to be checked vs. SCT reliability............................... 976-15 Schedule compression effect vs. SCT reliability............................................................... 987-1 Computerized dynamic project management model........................................................ 105A-1 Generic construction process structure under schedule compression.............................. 107A-2 Initial work rate structure................................................................................................ .108A-3 Work quantity accumulation structure............................................................................. 109A-4 Latency modification structure........................................................................................ 110A-5 Schedule pressure and work rates structure..................................................................... 111A-6 Iteration work rate and change and error structure.......................................................... 112A-7 Work release rate structure..............................................................................................113A-8 Workforce and productivity structure..............................................................................114A-9 Resource constraint structure........................................................................................... 115A-10 Dependency structure...................................................................................................... .116

PAGE 11

11 LIST OF DEFINITIONS Constraint Anything that lim its a syst em from achieving higher performance versus its goals Constraint thoroughness The pro cess to identify constrains de rived under schedule compression and to analyze their impacts on project Constraint management The process to stabil ize work environment by managing constraints and increase the accuracy of prediction and decrease uncertainties to create the best performance Evolution The rate at which design inform ation is generated from the start of an activity through the comp letion of the activity Latency The possibility of any failureerrors, changes, and side effects under schedule compression LLT Latent Lazy Time. The possibility of time reduction by applying schedule compression Latency thoroughness The process to detect latency generated under schedule pressure and manages to solve or reduce the problems Parkinsons Law Work expands so as to fill the time available for its completion QM Quality Management. Any actions to ensure all activities requirement and to improve their work performance Reliability The degree to which LLT could be detected through SCT Schedule compression A reduction of time availabl e to complete the work compared to the normal experienced time or optimal time for the type and size project being planned within a given set of circumstances SCM Schedule Compression Management. The process that validates the possibility of time reduction by managing LLT in accordance with schedule compression SCT Schedule Compression Thoroughness. The process to monitor and discover LLT including all unintended LLT Sensitivity The amount of work required of upstream information change Stability The degree to which the sche dule compression within the detected LLT scope would be performed successfully

PAGE 12

12 Yerkes-Dodson Law An empirical relationship between arousal and individual performance, which dictates that performance increase with cognitive arousal but only to a certain point 90% syndrome A project reach es about 90% completion according to the original project schedule but then stall; fina lly finishing after about twice the original project duration has elapsed

PAGE 13

13 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DYNAMIC PROJECT MANAGE MENT METHODOLOGY: MANAGING SCHEDULE COMPRESSION By Jaesung Lee August 2008 Chair: Ralph D. Ellis Major: Civil Engineering Despite aggressive efforts of project managers to maintain a project schedule or to recover from a lapsed schedule, delays and cost ove rruns have become routine phenomenon at many construction projects. This research proposes a proactive schedule compression method to reduce the expected project completion time by removing latent lazy time caused by fallible scheduling and constraints that impose non-value-added effect s on project. An additional objective of this research is to develop a systematic environm ental model, the Dynamic Project Management Model under Schedule Compression (DPM), to improve performance against latency and complexities of design and construction project s in schedule compression. Developing and implementing the DPM model will re sult in the following research outcomes: the association of schedule compression and project components with the constructi on process and their effects on construction performance; detecti ng and eliminating latent lazy tim e that project has potentially; managing latency caused by schedule compression; understanding of schedule compression frameworks; and developing system dynamics-based schedule compression model to improve performance under the condition of latency and comp lexities of design and construction project.

PAGE 14

14 CHAPTER 1 INTRODUCTION The project ma nagement team is responsible for finding methods of meeting the control budget and schedule rather than justif ications for not meeting them. in E/C contractors project management policy Delays and cost overruns are th e rule rather than the exception in construction (Sterman 1992). Working under schedule pressure and in a stressful environment has become a routine phenomenon at many construction sites (CII 1989; Nepal et al. 2006). Despite the aggressive efforts to maintain the project on schedule or to recover from a lapsed schedule, project managers walk the via Dolorosa. Time conservation is a prime concern fo r the both of owners and contractors on construction projects due to the fact that time inevitably equals money; consequently time savings greatly improve profits, while a loss of tim e can lead to financial distress (Chang et al. 2005). There have been many researches devoted to avoiding the losses of time and money by delay and developing models to control operatio ns and resolve the pr oblems. The efforts, however, have had limited success in terms of dealing with delay prev ention or keeping the schedule on time. Studies have also been conduc ted to determine how to pressure schedule effectively by means of increasing work hours and adding resources and how to manage labor productivity or disputes on the schedule pressu re. These studies suggested the alternative and passive solutions because the solutions are acte d after delay already takes place and demand to accelerate a project by aggressively pushing the schedule and to cost owner and contractor is necessary. Although accelerating a project can be rewarding, the conse quences can be troublesome (Thomas 2000; Pena-Mora and Park 2001; Nepal et al. 2006): productivity and quality are often sacrificed for the sake of remaining ahead of schedule (Ballard and Howell 1998b), and the

PAGE 15

15 negative effects of schedule pressure arise mainly from working out of sequence, generating work defects, cutting corners, and losing the motivat ion to work (Nepal et al. 2006). This implies, therefore, that the strategies de vised so far cannot be the optimal solutions for time conservation. Another important factor to be considered is the fact that cost and time are commonly estimated based on historical data and past expe riences of contractors and project managers in the planning and designing phases. However, because many projects have been struggling with delays and cost overruns, it is true that the historical data include the projects that did not finish on time and on budget. Furthermore, project perf ormance prediction relies on project managers judgments. These judgments are subjective and ambi guous decisions that will be made to control an urgent situation on the job site. For this r eason, there is a need for a better proactive and objective system into project control. The term schedule compression is defined as a reduction of ti me available to complete the work compared to the normal experienced time or optimal time for the type and size project being planned within a given set of circumstances (CII 1994). In spite of its important role in construction scheduling, only a small amount of research has been conduc ted on the effects of schedule compression and, more importantly, there are the li mited understanding on pressing schedule after delay and disrupti on, as mentioned above. In an e ffort to address this issue, researchers must recognize a timely and proactive manner which to mitigate any negative impacts on both schedule and cost is required. In addition, rather than historical data and subjective judgment, which causes un predictable results, more system atic and objective efforts to estimate cost and time in the condition of schedule compression must be analyzed.

PAGE 16

16 Figure 1-1. Proactive mannered scheduling This research is based on the assumptions th at there is an optimal pressed schedule at which performance is maximized. A project contains either potentially reduc ible duration that it originally consists of or preventable delays created by constraints that possibly can be mitigated or removed. As illustrated at (a) in Fig.1-1, a proj ect may have reducible duration at the planning phase. By the normal execution of construction, the duration can be re moved at the end of construction phases, (b) in Fig.1-1. Otherwise, the projected schedul e is accelerated ahead; such as (c) early schedule compression, (d) late schedule compression, and (e) split schedule compression in Fig.1-1. As a result, the executed dur ation can be shortened as much as reducible duration and thus, the need of proactive mannered scheduling is explained.

PAGE 17

17 Relying on the theories supporting these assump tions, this research intends to propose a dynamic project management model (DPM) to understand how schedule compression and constraints that impose non-value-adding effects on projec t are associated with the construction process and how they affect construction performance. Also, this research will develop the systematic environmental model in order to im prove performance under the condition of latency and complexities of design and construction proj ect. Thus, the proposed m odel is expected to benefit the entire life cycle of design and construction projects by reducing costs and duration, avoiding delays, increasing qua lity, and improving project management by eliminating counterproductive activities. The objective of this st udy is as follows: Objective 1: to identify latent lazy time and latency on construction process and their effects; Objective 2: to understand the association of project components with construction process under schedule compression; Objective 3: to develop systematic environmenta l dynamic model to manage construction project under schedule compression; Objective 4: to improve performance under high pr essured schedule by managing latency; and Objective 5: to validate the results by conducting case studies.

PAGE 18

18 CHAPTER 2 RESEARCH APPROACH 2.1 Research Assumptions The application of theory in psychology can elucidate how individual perform ance is related to stimulation, time, stress, and motivation. An individuals performance influences the entire process of a project and its results. That the time allotted for a task and an individuals performance of which tasks are interrelated is supported by two main psychological theories: the Yerkes-Dodson Law and Parkinsons Law. 2.1.1 The Yerkes-Dodson Law The relationship between arousal and work perform ance is curvilinear in which the optimum level of performance is obtained at an intermediate level of arousal (Wickens and Hollands 2000). The Yerkes-Dodson Law (Fig.2-1) explains an empirical relationship between arousal and individual performance modeled as the converted U-shap ed curve. It dictates that performance increases with cognitive arousal, but only to a certain point. Performance, however, Figure 2-1. The Yerkes-Dodson Law (adapt ed from Wickens and Hollands 2000)

PAGE 19

19 decreases when levels of arousal become too high. An inference is that there is an optimal level of arousal at which performance is at a maximu m. The upward portion of the converted U can be thought of as the energizing effect of arousal. Ne gative effects of arousal on cognitive processes on the other hand cause the downward portion. According to the Yerkes-Dodson Law and in relation to project management, we assume that there exists an optimum level of schedule pr essure at which performance is at a maximum. When schedule pressure is too low, the performan ce is affected because of a lack of urgency or awareness or through boredom, for example. On the other hand, when there is too much pressure, the expected performance may be difficu lt to achieve as a result of phenomena such as information filtration and omission, adaptation, frustration, and decreased human judgment, and coping strategies tend to be active (Nepal et al. 2006). In this reason, the Yerkes-Dodson Law indicates the importance of finding the optimal schedule pressure to maximize the product performance and explains that appropriate and well managed sche dule pressure can increase their productivity up to a particular level (Lee et al. 2006a). 2.1.2 Parkinsons Law Parkinsons Law states that work expands so as to fill the tim e available for its completion. It can be explained as follows: the demand upon a resource always expands to match the supply of that resource. Put another way, it is if you give someone 8 hours to do a 2-hour project, it will take the full 8 hours to get done, not 2 hours. In Fig.2-2, the planned work, the area of p1 x t1, is equal to the complete work, p2 x t2. Parkinsons Law supports the assumption that there exists the late nt lazy time which creates rubber band duration. Time added to the original duration may not effectively protect the planned schedule because when people realize that they have more time to complete a task than the time actually specified, their work productivity usually goes down, often with the task being

PAGE 20

20 Figure 2-2. Parkinsons Law (modi fied from Lee et al. 2007) deferred to the last minute. This phenomenon is explained in 90% syndrome as well. 90% syndrome defines a project that reaches ab out 90% completion according to the original project schedule but then stalls; finally finishing after about twi ce the original project duration has elapsed. Ford and Sterman (1999) found out th at the average time to develop a product was 225% of the projected time, with a standard devi ation of 85%. The study, furthermore, insists that the management of the final 10% during the last half of the project is typically of the most concern to managers because this portion of th e project obviously and significantly deviates from planned progress, focusing attention on the failure of the project to meet its targets, and that eliminating the 90% syndrome could potentially reduce cycle time roughly 50% (Ford and Sterman 1999). 2.2 Research Applications Based on the Yerkes-Dodson Law, the optimal schedule pressure exists at a m aximum of performance. When schedule pressure is higher or lower than the optimal schedule pressure, it induces lower performance and the lower performa nce results longer durati on than expected. The

PAGE 21

21 optimal schedule pressure supported by the Yerk es-Dodson Law will improve the performance. And this affects the relations hip between the project duration and the schedule compression. Putting more pressure on the schedule can resu lt in the time reduction for a project. When the time reduction can be detected by schedule pressure and the performance remains the same amount as it was (t2 t1), the time difference ( t = t2 t1) between the expected duration (t2) and the pressed duration (t1) indicates latent lazy time (Fig.23). This statement is supported by Parkinsons Law, which assumes there may be more time allocated to complete a project than is actually needed. If this latent lazy time can be detected and eliminated or reduced by the imposition of schedule pressure, the reduction of total project dur ation will result. The Yerkes-Dodson Law proposed that different tasks might require di fferent levels of arousal. Difficult or intellectua lly demanding tasks may require a lower level of arousal for optimal performance to facilita te concentration, whereas tasks demanding stamina or persistence may be performed better with higher levels of arousal to increase motivation. Figure 2-3. Time reduction by eliminating latent lazy time (Lee et al. 2007)

PAGE 22

22 Figure 2-4. Worse performance by sche dule pressure (Lee et al. 2007) When there is no latent lazy time or schedul e pressure is higher or lower the optimal schedule pressure, schedule pres sure causes worse performance (p1 p2) and it affects the project duration (t1 t2); A B (Fig.2-4). This is the last objec tive of this proj ect. If, under schedule pressure, we can determine the reason causing worse performance, called latency, we can institute strategies to improve work perf ormance under high schedule pressure without delay. Further, we should consider the preparati on about unexpected results on the high schedule pressure. 2.3 Methodology Overview To accomp lish the objectives described in th e introduction, this research was conducted based on three major steps: analysis, development, and validation. In the an alysis step, research objectives were established and research compon ents to be studied were determined through diverse methods. Based on this analysis, the components were evaluated and applicably developed to the research objectives by syst ematic concepts and models: constraints management and system dynamics. Then, the models and all components developed in the

PAGE 23

23 Figure 2-5. Concept of dynamic project management model (Lee et al. 2007) previous stage were validated through a couple of real-world case projects. It is important to note that these steps are not sequential, but spiral and parallel at some parts. To satisfy the research objectives, first two main studies, project management under schedule compression and the improvement of re liability and stabilit y of projects, were conducted simultaneously. As mentioned before although accelerating a project can reduce project duration and satisfy owners requirements, productivity and quality are often sacrificed and the actual benefits may not be worth the ti me saved. By these reasons, the reliability and stability of projects are pr erequisites to the implementa tion of schedule compression. For reliability and stability to be improved, it is n ecessary to identify and manage the constraints creating non-value adding activit ies. Constraints management a nd lean construction play the critical role in the improvement of reliabilit y, rather than the reduction of project duration. Consequently, the proactive ma nnered strategies controlling th e constraints should be set up simultaneously with the schedule compression (Fig.2-5). 2.3.1 Analysis In this analysis step, diverse research me thods literature review, practitioner interview and questionnaire survey were c onducted to identify what to solve and how to solve in detail. Dynamic Project Management Model Management of Schedule Compression Improvement of Reliability and Stability

PAGE 24

24 Figure 2-6. Research methodology According to the data relied on the diverse methods the identification and the influence of latent lazy time and latency, and their association w ith construction process were analyzed, and the constraints deteriorating reliability and stability were identified. Based on the analysis, major issues for the development were determined. 2.3.2 Development There are three m ajor development phases: development of systematic frameworks for schedule compression, constraints management framework, and the combined management mechanism based on system dynamics approach Through the phase of development of the schedule compression framework, mo re practical and realistic concepts and methods were ANALYSIS VALIDATION Case Study Policy Analysis DEVELOPMENT Systematic Framework Under Schedule Com p ression Systematic Framework for Constraints Mana g ement Management Mechanism System dynamics-based Com p ression Model Define Latent Lazy Time and Latency and their Effects Identify Components affecting Reliability anditsEffects

PAGE 25

25 determined based on the results of the analysis step. At the same time, a constraints management model was established to improve reliability and stability of projects. The analyzed frameworks aim to understand how the influen ce of schedule compression is associated with constrains, and how they affect construction performance. Finally, based on the frameworks, management mechanism was created to manage the detrimenta l impact of these components and simulated via system dynamics-based model. 2.3.3 Validation As the last step, all developed dynami c project management models are applied to schedule compressed activities of real-world case projects to be verified. Through this process, the result of validation shows how close th is model presents to the actual performance. The validation phase is conducted through three steps: input elicitation, applica tion to DPM, and comparison of model value with actual value.

PAGE 26

26 CHAPTER 3 SCHEDULE COMPRESSION MANAGEMENT FRAMEWORK According to the Construction Industry In stitute (CII 1990), schedule compression is referred to as the shorteni ng of the required time for accomplishing one or more engineering, procurement, construction or startup tasks (or a to tal project) to serve one of the three purposes: (1) reducing total design-constr uction time from that considered normal; (2) accelerating a schedule for owner convenience; and (3) resolving lost time after falling behind schedule. Also, CII states that the primary reasons for compre ssing or accelerating the sc hedule of a construction project can be attributed to the following: (1) monetary considerations such as project financing, lost producing during constructi on, or stockholder pressure; (2) the development of a new product or service by the owners organization that needs to get to market as soon as possible due to rising loss-of-opportunity costs; and (3 ) the planning and design phases of the project delivery cycle have fallen behind the required schedule, forcing the construction phase to make up the lost time (CII 1990; Noyce and Hanna 1998). In this phase, the association of construction process, latent lazy time, and latency under schedule compression and how they dynamically a ffect construction performance are discussed. The frameworks for this purpose consist of tw o main frameworks, the internal management framework and the external management framework, which are combined for the entire project management. These frameworks are the bases for the next step, the management mechanism, in order to develop system dynamics-based proj ect management model under schedule compression model. 3.1 Literature Review Schedule compression is commonly regarded as a tim e-cost trade-off problem between the amount of compression and the consequent increase in direct costs due to schedule compression

PAGE 27

27 (Yae et al. 1990; Noyce and Hanna 1997). Many heuristic methods and models have been developed as means of mathematically compress ing individual activitie s of a project schedule (Perera 1980; Perera 1982; Coskunoglu 1984; Ritc hie 1985; Vrat et al. 1986; Ritchie 1990; Yae et al. 1990; Moselhi 1993; Senouci et al. 1995; Noyce et al. 1997). Despite the attribution of the schedule compression to the project duration, very little has been published in regard to office and field techniques used to compress a schedule originally developed using normally expected durations. Only a few studies attempted the disc rete ways such as work force, financial incentives, overtime, and work scheduling. The notable studies for schedule compression ap plicable to office and field have been conducted by Construction Industry In stitute, in University of Texa s Austin. Based on the expert survey and interview, CII defined the 94 concepts and methods for schedule compression through Delphi methods. The study identified sche dule compression techniques that can be used in one or more of the engineering, procuremen t, and construction phases of the project, and evaluate each techniques impact on the cost and duration of the project wh en applied at the three phases of the project. Moreover, CII gives the definition on each techniques based on the expert survey and literature review. For example, th e definition of Just-in-Time Material Deliveries in Material Management is; 5.03 Just-in-Time Material Deliveries Deliver materials to the work place as they are needed without intermediate on-site storage. This technique eliminates the time normally allowed for on-site storage. However, successful execution of just-in-time deliverie s requires extreme planning, coordinating and expediting action since any failure in the process can produce delays throughout the system. A fringe benefit of this technique is the elimination of double handling of materials on site with consequent re duction of work-hour requireme nts and potential for loss or damage in handling or storage (CII 1990).

PAGE 28

28 Based on the CIIs schedule compression concepts, Noyce and Hanna delineated planned and unplanned schedule compression to increase pr oductive time and reduce the project schedule. This model contains 34 concepts and methods that are determined to most directly apply to the construction phase. Planned schedule compression is defined as schedule compression that was anticipated and planned for before the start of the construction phase of the project, whereas Unplanned schedule compression is defined as schedule compression that was not anticipated and planned for before the star t of construction. Unplanned sc hedule compression is commonly a result of some form of unanticip ated change to the originally planned scope of the work and/or construction schedule (Noyce and Hanna 1998). With a different approach, Nepal et al. (2006) analyzed the effects th at schedule pressure has on construction performance, and focuse s on tradeoffs in scheduling and developed scheduling strategies resolving the negative ripp le effects. A research framework has been developed using a causal diagram to illustrate th e cause-and-effect analysis of schedule pressure. This study indicates that the advantages of increasing the pace of workby working under schedule pressurecan be offset by losses in productiv ity and quality. The negative effects of schedule pressure arise mainly by working out of sequence, generating work defects, cutting corners, and losing the motivation to work. The adverse effects of schedule pressure can be minimized by sche duling construction activities realistically and planning them proactively, motivating workers, and by establishing an effective project coordination and communication mechanism (Nep al et al. 2006). Desp ite their tremendous works, the researches all failed to show how the effects of schedule compression work systemically to construction process. Accordingl y, the systemic understanding and analysis about schedule compression are required.

PAGE 29

29 3.2 Feedback Process of Schedule Compression Construction is inherently dynam ic and involve s multiple feedback processes that produce self-correcting or self-reinforci ng side effects of decisions (S terman 1992; Pena-Mora and Park 2001). These feedback processes contribute to ge nerating indirect and/or unanticipated events during the project execution and make the cons truction process dynamic and unstable, which cannot be captured in the traditional planning tools (Park 2001). Dynamics in a system arise from the intera ction of two types of feedback processes, reinforcing and balancing, among th e components of the system, not from the complexity of the components themselves (Sterman 2000). Fig.3-1 show s the conceptual reinforcing and balancing feedback process of latent lazy time and la tency by schedule compression. Reinforcing loops tend to reinforce or amplify whatever is happe ning in system. Those are all processes that generate their own growth. Balancing loops coun teract and oppose change, which describes all processes that tend to be self -limiting and that seek balance and equilibrium (Sterman 2000). Under schedule compression, projects can generate two effects; time reduction or side effects. Time reduction can be achieved by removing latent lazy time. And, ill-managed schedule compression can cause side effects such as errors and changes creating delays or reworks. The balancing loop shows that the management to re ctify latency and reduce latent lazy time can have the intended effect of reso lving the issues if the decision is correct and well implemented. At the same time, the reinforcing loop illustrates that it can produce side effects that may augment some unintended problems, such as if the decision is incorrect, not well implemented, exceeds the time frame of its effec tiveness or if a project manager does not realize the impact of the control actions on other related activities.

PAGE 30

30 Figure 3-1. Conceptual feedback loop under sche dule compression management (Modified from Lee 2006) 3.3 Latent Lazy Time and Latency Before project launch, anticipat ing the exact process of it is truly hard. Contractors and project ma nagers commonly estimate costs and ti me based on the historical data from past projects that they have done. Many projects, however, have been struggling in delays and cost overruns. Hence, it is true that th e historical data includes the pr ojects that did not finish on time and within budget. In other words, there could be a time that can be reduced on execution as seen in Fig.1-1. Latent lazy time is define d as the possibility of time reduction by applying schedule compression. If optimal duration is shorter than planned duration, the difference between the former and the latter is latent lazy time. The reducible durati on in Fig.1-1 and the 6 hours in the example of Parkinsons Law, which is that it w ill take the full 8 hours to get done not 2 hours if you give someone 8 hours to do a 2-hour proj ect, all can be latent lazy time.

PAGE 31

31 According to the Yerkeys-Dodson Law, there exists an optimum duration on the level of schedule pressure at maximum performance. Wh en schedule pressure is too low or too high, however, the expected performance may be difficu lt to achieve as mentioned before. That could result phenomena such as failure, errors, changes, or delays. These sort of side effects are called latency (Fig.3-2). Latency is defined as the possi bility of any failuree rrors, change, and side effects with or without schedule compression. Latent lazy time can be e xplained by the results-oriented control method. The resultoriented control is intended to reveal problems so they can be solved. Suppose a compression concept, Just-in-Time (JIT) material deliveries, is applied to an activity. As described in literature review, JIT concept can eliminates double handling of ma terials on site, so that the technique has the benefit of consequent reducti on of time allowed for on-site storage and workhour requirements, and potential for loss or damage in handling or storage (CII 1994). The time reduction after accelerating schedule, JIT in this ex ample, can be defined as latent lazy time. We can detect the latent lazy time and reduce time and cost by this method, which is also assumed and depicted in Fig.2-3. Figure 3-2. Effect of latent lazy time and latency (Lee et al. 2007) Latent Lazy Time Latency Schedule Compression Activity Side Effects Time Reduction

PAGE 32

32 On the other hand, schedule compression may result site effects (Fig.3-2). Back to the JIT example, if any failure in the JIT process follo wed by mal-coordination of implementation of illimplementation of the action occu rs, unexpected delay is produced through the system. In this context, the side effects after applying the co mpression concept are latency. Latency is divided into two components: existing latency and nonexisting latency. Existing latency is any possibility of failure such as unforeseen errors and changes that projects inherently have. Nonexisting latency is the site effects after appl ying schedule compression such as time delay after JIT in the previous example. 3.4 Relationship of Activities Traditional project managem ent methodologies based on the Critical Path Method (CPM) and the Project Evaluation and Review Techni que (PERT) describe the relationship of construction processes and activities as static and linear. However, the actual relationship of real world is more dynamic and complicated. For closer to the real model, the concepts of evolution and sensitivity are applied to this project. 3.4.1 Evolution and Sensitivity Krishnan et al. (1997) insisted that the information from the upstream to the downstream and the dependency of the downstream to the in formation from the upstream are the important components of product development to product fast er, especially in the overlapping process. Evolution is referred to as th e refinement of the upstream information generated from its preliminary form to a final value in their projec t (Krishnan et al. 1997). Also, Bogus et al. (2005) define evolution as the rate at which design information is generate d from the start of an activity through the completi on of the activity.

PAGE 33

33 Figure 3-3. Information dependency Therefore, project launch highly relies on th e completeness of the information of planning and design phases according to the dependenc y of them. In the f oundation activity, pouring concrete cannot be started before form work has been done fully or partially. Such a relationship connecting activities is information dependenc y. The downstream activity is affected by the information from the upstream activity. So, evoluti on entails how fast to release the information from the upstream to downstream. Evolution is identified by de sign optimization, constraint sati sfaction, internal and external information exchange, and standard ization (Fig.3-4). A faster evol ution does not have to mean a shorter duration for an activity. Evolution is define d in terms of the rate at which information is generated, which does not necessarily relate to the overall duration of an activity. However, activities with fast evolution are more amenable to overlapping than activities with slow evolution. Removing information dependencies allo ws sequential ac tivities to reduce the project schedule (Bogus et al. 2005).

PAGE 34

34 Figure 3-4. Determinants of evolution and sensitivity Sensitivity is the amount of rework required of upstream information change (Bogus et al. 2005), and it is determined by constraint sensitivity, input sensitivity, and integration sensitivity (Fig.3-4). For example, installing pipeline is highly sensitive to pouring conc rete. If the pipelines are not properly installed, the concrete foundation has to be removed and the work of pouring concrete has to be redone. Moreover, this rela tionship will affect the entire project duration. Identifying the sensitivity toward downstream activities is important in planning decision. Starting a highly sensitive activity before all upstream information is complete entails an increased risk that significant rework will be required. The faster the evolution of information in an activity, the less ris ky it is to begin a downstream activity before the upstream activity is finalized. Also, the lower the sensitivity to changes in upstream information, the less risky it is to overlap activ ity (Bogus et al. 2005). 3.4.2 Dependency Ford and Sterma n (1998a) researched how dyna mically the role of these relationships plays in the construction progre ss. Fig.3-5 depicts the internal dependency in situations with constraints affecting work progre ss (A) and without constraints (B). It can explain that in the situation on B the reliability of latency is 0%. So, the possibility of success of schedule compression would be 100% and time reduction will be expected as much as latent lazy time.

PAGE 35

35 Figure 3-5. Internal dependenc y (Pena-Mora and Park 2001) Constraints, however, exist in most construc tion processes and can affect the development progress and time of the construction. Graph A in Fig.3-6, which is about the external dependency, explains the relationship that the downstream activity is scheduled to start at 50% completion of the upstream activity, and then finishes with the upstream. It means that the downstream work is dependent on the upstream activity partially. Graph B represents that th e downstream activity can start and finish Figure 3-6. External dependenc y (Pena-Mora and Park 2001)

PAGE 36

36 regardless of the process of the upstream activity. In this case, two processes are not related to each other. In this context, applying these concep ts, the internal and external relationships allow anticipating and managing work proce ss more practically and efficiently. 3.5 Schedule Compression Management Framework At the project design and planning phases, project duration and cost are estimated based on the historical data from the completed project to achieve the accura cy of prediction. The characteristics of construction pr oject, however, illustrate its la rge size, uniqueness, uncertainty, on-site production, and highly dyna mic complexity, and these char acteristics abate certainty of prediction, and the worse prediction is not able to manage construction process effectively. Latent lazy time and latenc y are hard-predicted and hard -managed variables by their characteristics of unfamiliar identification and co mplex process. Maybe, they could be detected and analyzed even after completion of projects. For this reason, the systematic identification and understanding of the construction process with respect to latent lazy time and latency under schedule compression in terms of project and activity level is mo stly required to manage the variables successfully. The multiple feedback processes in construction project necessitate that LLT and latency must be considered and managed as continuous and constant factors. One response to the variables could generate other responses in iterative cycle. The scheduled compression management framework helps to identify th e generation and management of iterative relationship focused on LLT and latency under schedule compression. The frameworks are supposed to be proposed in the way of inte rnal management and external management framework. The holistic view of frameworks facil itates to draw the big picture of project and identify the inter-connection of its components.

PAGE 37

37 3.5.1 Internal Project Management Framework As defined before, LLT is the possibility of time reduction by m anaging project schedule systemically. When planned project duration is larger than the optimal duration and does not have any intention such as time buf fering at the start stage, the proj ect originally involves LLT. It just comes from the nature of project derived fr om the historical data during the estimation of project duration. The internal project manageme nt framework scrutinizes the relationship and effects of LLT and latency with other com ponents under schedule comp ression within a single activity. All tasks and activ ities should be monitored and addr essed to detect LLT before the judgment of the possibility of schedule accelerati on. In order to depict this process, Schedule Compression Thoroughness (SCT) is introduced. SCT is the process to monitor and discover LLT including unintended LLT and the unintended LLT is generated from upstream activities or LLT reevaluation after the failure of SCT. The term, reliability is used in this step to indicate the degree to which LLT could be detected through SCT. For example, if the estimated duration of an activity is 100 days, th e predicted LLT of the activity is 10%, and the activity has 80% reliabil ity on SCT, then the total reducible time by LLT is 10 days, and 8 days out of 10 days are detected through SCT. It is considered the actual time that can be reduced by schedule compression within the total work scope of the activity. These relationships are formulated as follows; LDDPL Eq. 3-1 RDDLA Eq. 3-2 So, )( RLDDPA Eq. 3-3

PAGE 38

38 Figure 3-7. Internal management framework

PAGE 39

39 where, DP is the planned duration, DL is the total reducib le duration by LLT. DA is the actual reducible duration through SCT, L is LLT, and R is reliability. However, it is impossible to identify all of LLT through SCT. Therefore, the unidentified LLT should go back to the monitoring step again, and this step is iterative until LLT is completely removed. By this means, L and R are reformulated. n k kLLL2 1 Eq. 3-4 n k kRRR2 1 Eq. 3-5 Detected LLT at the first attempt is determined by L1 and R1. Others are detected by n k kL2 and n k kR2. However, the completed detection of LLT is impossible in real world, and the LLT and reliability on SCT can vary due to the impact of a diverse set of variables during the work process. So, an alternativ e step is needed to supplement these deficits. This will be explained in the step of Quality Management process. The identified LLT steps forward Schedul e Compression Management (SCM) to be managed and removed. But, unidentified LLT is remonitored with unintended LLT before the work scope is adjusted. SCM is the process th at validates the possibi lity of time reduction by removing LLT in accordance with scheduled compression. With the known characteristics of schedule pressure, it is very difficult to execute time reduction without any loss of perf ormance from side effects, such as out-of sequence work, cutting corners, losing motivation, fatigue, and erro rs and changes. So, the role of SCM reducing project duration by eliminating LLT in stable wo rk environment is the most important, along with consideration of proper tec hniques for schedule compression.

PAGE 40

40 To satisfy this condition, the concept of stability is brought in with the SCM process. Stability designates the degree to which the execution of schedule compression within the detected LLT scope would be performed successf ully. High stability of schedule compression increases the success of execution, and it results in more time re duction. In the application of schedule compression in SCM, for instance, an activity has 50% stability on schedule compression means 50% possibility of time reduc tion from the detected LLT in the previous example for SCT. Hence, the 50% stability of schedule compression process creates time reduction by as much as half of detected LLT from SCT, 4 days in the example, and that would be the final actual reduced dur ation through SCM process. Then, the executed project duration would be 96 days by 4 days, shortened from the planned duration. In the formulation, SDDAF Eq. 3-6 then, PFTDDD Eq. 3-7 where, DF is the final actual reducible duration th rough SCM, S is the stability for schedule compression on an activity, and DT is the total executed duration of project. Evolution is the dependency of information for work process before the execution as described fully in Chapter 3.4. The information of identified LLT by the reliability through SCT and the stability for schedule compression application decided in the SCM process influence the decision of implementation and its management during project process in accordance with the information of upstream work process, downstr eam activity readiness, constraints, design completion, and so forth. Only high evolution enables works stability.

PAGE 41

41 Then, all of these factors decide how singl e activity is dependent on each other. For example, in the foundation activity, pouring concre te work totally relies on the completion of formwork as seen earlier. If reducible time (LLT) is detected in formwork, project managers have to prepare and execute schedule accelera tion. But, the stability of removing LLT on the formwork significantly makes the influence on th e time and the process of the concrete pouring work. Like this, the upstream information releas e and the downstream readiness and steadiness in an activity are the internal evolution and sensitivity, and they are thoroughly managed through the SCM process. These tasks are all involved in SCM process and increa se the creation and the completion of schedule compression in the model. If the schedule compression is well-managed and LLT is successfully removed from the first work scope, the result will generate the time reduction and the work scope will be adjusted according to schedule change, and then the activity is released to the downstream activities or the next phase of project. However, if the situation under schedule comp ression is not managed effectively or other unforeseen problems hinder the work process du ring the implementation of an activity, some Figure 3-8. Evolution loop

PAGE 42

42 portion of errors or delays, such as unintended side effects, can be created. These side effects are not predicted under the normal project process, which is the planne d work process. For instance, the failure of schedule compression, which would be out-of sequence work by workers fatigue or misunderstanding of LLT by project managers, genera tes derivative errors and reworks. But, it will not happen if the schedule is not accelerated. In terms of this, these side effects are called latency as described before. Latency thoroughness is the process to detect latency generated unde r schedule pressure and manages to solve or reduce it and its problems the way the f unction of the processes of SCT and SCM works. Quality Manageme nt (QM) is defined as any ac tions to ensure all activities requirement and to improve their work performan ce. For this, all tasks about the management of latency belong to the QM process. Then, the ma naged side effects that may include other LLT should be monitored again and take the iterative steps to manage LLT and latency. Additionally, the information from the internal evolution to QM, the shadowed area, affects decisions and influence for SCM to manage LLT under schedule compression. Consequently, the development of informati on on upstream activity increase the evolution, evolution increases the predictabi lity about the situation of work s, and then the predictability finally stabilizes the execution of the downstr eam or the activity under schedule compression. In addition, high stability ameliorates the correct ness of information to downstream (Fig.3-8). Adjusting scope is used to address the total im pacts on projects by the processes of internal project management framework and introduce new sc ope of work to be controlled at the next steps. If the scope of work or project is not adjusted according to the management of schedule and works, it causes an overflow of work and c onsequently, could cause a project to suffer from slow progress at the later steps (Lee, 2006).

PAGE 43

43 Figure 3-9. External management framework

PAGE 44

44 3.5.2 External Project Management Framework The external project ma nagement framework in Fig.3-9 explains the impacts of schedule compression managing LLT and latency on activities to other related activities of construction processes. The relationship between the upstream activit y and the downstream activity is connected by the concepts of sensitivity and evolution of each activity. The SCM and QM processes of upstream activity influencing its own SCM process, the shadowed area, provide information to downstream as well. This helps manage LLT on the downstream activity, and this process is defined as the external evolution. Figure 3-10. Extended project management framework

PAGE 45

45 LLT existing in the downstream activity is sensib le to external rela tionship of activities. Unintended LLT, for example, is originally generated from external environment. And, unmanaged latency from upstream also may create unpredicted LLT to downstream. On this account, the LLT of downstream is influenced by the total impact of th e upstream activity. Even though the external management framewor k starts from the simple interrelationship by sensitivity and evolution between upstream and downstream, the real work process is extended and much more complicated and dynamic (Fig.3-10).

PAGE 46

46 CHAPTER 4 CONSTRAINTS MANAGEMENT FRAMEWORK Goldratt (1988) defined a constraint as anything that limits a system from achieving higher performance versus its goals and insisted that ev ery system must have at lease one constraint. Whether it is originated from in a project or is derived from factors of project processes, constraint is the target that must be iden tified and managed to improve performance and productivity of a project. As mentioned before, although accelerating a project can reduce project duration and satisfy owners requirements, productivity and qualit y are often sacrificed a nd the actual benefits may not be worth the saved time. Hence, for th e success of the schedule management, reliability and stability of project s are prerequisite, constraints creating non-value adding activities should be managed, and the proactive manner strategi es controlling the constr aints should be set up simultaneously with the schedule compression. As seen in Fig. 3-2, accelerating schedule could either reduce project duration or generate side effects. The objectives to be obtained in th is phase are to improve re liability and stability by reducing uncertainty and variation, thereby providing a better envi ronment, enhancing the effects of applied schedule compression and reducing co mpletion time in a project. The ways to improve stability include increasing predictabil ity of downstream activities, removing insecure variables and wastes, decreasing sensitivity by less dependency, reducing uncertainties, and managing constraints. For this constraints manage ment framework, the theory of constraints, key constraints management, and shielding producti on system based on lean construction were applied to create constraints management framework and satisfy the prerequisites.

PAGE 47

47 4.1 Applied Concepts 4.1.2 Theory of Constraints Based on the Goldrattes concept (1988), Ra hman (1998) summarized the Theory of Constraints (TOC) as: Every system must have at least one constraint ; and the existence of constraints represents opport unities for improvement. Constrai nt is the weakest link among activities or bottleneck in a pro cess. Any improvement in the cons traints performance translates directly into improved overall system performance. By strengthening th is weakest link, the whole process is stronger; by increasing the flow through the bottleneck, ove rall system output is increased (Womack and Flowers 1999). Due to th e limitation of resources and time, however, it is impossible to explore every sing le constraint in the system fo r the best result. For practical resolution, it is necessary to loca te the most crucial constraints and resolve them with the highest priority, according to which resolving the constraints at the bottleneck production processes leads to enhanced overall system performan ce. By these reasons, Goldratt (1990 and 1992) proposed a five-step generic procedure fo r ongoing improvement in the TOC (Fig.4-1): (1) Identify the systems constraint(s). Identify these constraints and also necessary to prioritize them according to their impact on the goal(s) of the organization. (2) Decide how to exploit the systems constraint(s). A managerial cons traint should be eliminated and replaced with a policy, which will support increased throughput. (3) Subordinate everything else to the above decision. Every other component of the system (nonconstraints) must be adjusted to support the maximum effectiveness of the constraint. Because constraints dictate a firms throughput, resource synchronization with the constraint provides the most e ffective manner of resource utili zation. If nonconstraint resources are used beyond their productive capacity to support the constr aint, they do not improve throughput, but increase unnecessary inventory.

PAGE 48

48 Figure 4-1. Process of on-goi ng improvement (Rahman 1998) (4) Elevate the systems constraint(s). If existing constraints are st ill the most critical in the system, rigorous improvement efforts on these constraints will improve their performance. As the performance of the constraints improves, th e potential of nonconstrai nt resources can be better realized, leading to improvement s in overall system performance. (5) If in any of the previous steps a cons traint is broken, go back to step 1. 4.1.2 Key Constraints Analysis With possibly hundreds of constraint s in existence at a tim e in th e course of a project, it is not realistic to consider all constraints equally important gi ven the fact that both time and resources are limited. Therefore, Goldratt suggest th at it is more efficient to improve the overall system performance through controlling constraint s at the bottlenecks and iterating the procedure for incremental improvement.

PAGE 49

49 A bottleneck is formed when the rate of re leasing job orders exceeds the capacity of a machine. In most cases, bottlenecks occur only at a few places while the production system, in general, may still have extra capacity. Bottleneck s exist when there are impediments in the flow caused by unresolved constraints so that work is not released fa st enough for downstream activities. It makes sense that constraints at th e bottleneck should be highlighted and tackled with priority, especially those resulting in delays (Chua el al. 2005). Such constraints are denoted as key constraints. 4.1.3 Shielding Production Ballard and Howell (1998b) ma intained that the construction produc tion control system should be erected in terms of production planning, mate rial coordination, and work management, and the system should be considered in four di fferent project levels, in itial planning, lookahead planning, commitment planning, an d methods planning. Initial pl anning produces the project budget and schedule, and provides a coordinating map that pushe s completions and deliveries onto the project. Lookahead planning details and adjusts budgets and schedules and pulls resources into play, thereby focusing supervisors attentions toward what is supposed to be done in the near future as well as directing their pr esent actions in a way ensuring that the desired future actions occur. Commitment planning is a commitment to what will be done, after evaluating should against can based on actua l receipt of resources and completion of prerequisites. And planning methods decide how work is actually going to be done with more detailed specification of methods from top to bottom (Ballard and Howell 1998b). Among these planning levels, lookahead and co mmitment planning enable the system to specify desired input and process so possibly anticipate output a nd decrease uncertainties within projects.

PAGE 50

50 Figure 4-2. Shielding producti on (Ballard and Howell 1998b) Shielding production (Fig.4-2) is an alternat ive strategy, which shields the direct work force from upstream variation and uncertainty. Sh ielding occurs from selecting only assignments that can be successfully completed, assignments for which all materials are on hand and all prerequisite work is complete. In order to avoid a mismatch between labor force and workflow, shielding is required to match labor and labor-re lated resources (tools, construction equipment, temporary facilities, etc) to the workflow into ba cklog. Lastly, in order to perfect the shield, the degree of fit between the works that have been d one and that will be done must be measured, the root causes of failures to complete planned work identified, and those causes to prevent repetitions eliminated (Ballard and Howell1994a ). Shielding is accomplished by making quality assignments, thereby increasing the reliability of commitment plans (Ballard and Howell 1998b). 4.2 Constraints Management Framework The purpose of the constraint ma nagement framework is to create the management structure with which to understand the associ ation of construction process and constraints generated by side effects under schedule pressure and to with which to manage the constraints (Fig.4-3).

PAGE 51

51 Figure 4-3. Constraints management framework Constraints thoroughness is the first step in identifying constraints derived under schedule compression and in analyzing their impacts on the proj ect. This step requires the iterative cycle to detect all possible constraints. Identified constraints are evalua ted and prioritized according to their criticalities to the project. The criticality of constraint is determin ed by its impact on project performance. Top ranked constraints through evaluation and priori tization process cause larger defects to the

PAGE 52

52 project rather than the others and make it more critical. By this mean, reducing and eliminating critical constraints are more practical and effective to achieve the required performance. After evaluation and prioritization of constrai nts, the reasons for delays or any other deficiencies caused by constraints should be explicitly determined based on the consideration of information, resources, and processe s involved in activities. This pr ocess leads to the necessity of adjusting dynamically work process, scope, and schedule in accordance with the changes of constraints. A stabilized environment is a prerequisite for a successful completion of a project. Only the stability of work environment produces th e rigid information flows from upstream to downstream and provides the improvement of downstream performance, so that the right sequence of work can move best toward project objectives (Fig.4-4). In order to achieve this objective, it is important to pr otect the direct work from upstr eam variation and uncertainty. A prediction system estimating the level of stability is introduced and facilitates the understanding of: (1) what characteristics of an activity or the project lead to defects or delay, (2) what the causes are, and (3) the understa nding of to understand how these causes are related to effects (Motawa et al. 2007). Through these st eps, only stabilized work is released for next processes. Figure 4-4. Generation from stabilized environment

PAGE 53

53 CHAPTER 5 SYSTEM DYNAMICS-BASED PROJECT MANAGEMENT MODEL The dynam ic and iterative processes of constr uction increase its complexity and this may balance the systematic structure of construction proce ss by itself or reinfo rce side effects of decisions. Network-based tools, such as Critic al Path Method (CPM) and Program Evaluation and Review Technique (PERT), do not explicitly c onsider the feedback relationships of activities and other soft variables affecti ng a project. A system dynamics approach provides an alternative view considering these major relationship and influences on a project and suggesting much of the detail to increase project performance with diffe rent ways that traditional project management tools do not employ. In the previous chapters, the frameworks were developed to analyze the association of construction process, latent lazy time, late ncy, and constraints under schedule compression and how they dynamically influence one another and affect construction performance are discussed. Based on these progressions, this chapter de velops a system dynamics-based project management model to simulate how schedule compression is managed, how this engenders malfunction, how they are combined with constraint management, and how they work in real environment. 5.1 System Dynamics System dynam ics was developed in the late 1950s to apply control theory to the analysis of industrial systems (Richardson 1985 ). Since then, system dynamics has been used to analyze industrial, economic, social, and environmenta l systems of all kinds (Turek 1995). System dynamics is a method for studying the world arou nd us. Unlike the tradi tional study by breaking object up into smaller and smaller pieces, system dynamics looks at things as a whole. The central concept of system dynamics explains how all the objects in a syst em interact with one

PAGE 54

54 another. The objects and people in a system inte rrelate through feedback loops, where a change in one variable affects other variables over time, which in turn affects the original variable, and so on. What system dynamics attempts to do is to understand the feedback of system and thus propose the behavior it can produce. The traditional project management tools experi ment with the data from a projected work before it is executed. From the data, the work is decomposed into elements that can be individually related to previous experience. It is then possible to produce reasonable estimates for each elements duration, cost, and resource requirements (Rodriques and Bowers 1996a). By this imposed discrete view, it is impossible to analyze and revise problems at the point where those happened. The system dynamics approach however, can capture the major feedback processes to solve defected components of a proj ect where detected until they have completely subsided. The other major roll of system dynamics model is the way to approach a project based on a holistic view. The strictly discrete view of the traditional tools may not be appropriate to the continuous nature of construction projects. Traditional tools can pr ovide a detailed description of the process including specific estim ates of costs and duration and detecting the direct causes of the impacts of project is possibl e. But, these methods are limite d by their use of an indirect project measure and by bundling th e characteristics of and relati onships among scope, resources, and processes in each activity in to a single duration estimate. Th ey also tend to ignore iteration or require that iteration be implicitly inco rporated into duration estimates and precedence relationships (Ford and Sterman 1998a). On th e other hand, the system dynamics approach considers highly aggregated views of project st ructure and the focus on the understanding at the project level instead activity level.

PAGE 55

55 Table 5-1. Comparison of the traditional approach and system dynamics approach Nature Traditional Approach System Dynamics Approach Point of View Approach Dependency Input Analysis Linear Progress Discrete and particular Focus on problem Precedence Resources What-if analysis Linear Fixed Continuous and holistic Feedback Internal and external dependencies in the entire project duration Resources and soft variables Policy analysis and guidelines Linear and Non-linear Varied The traditional models view the relationship am ong activities of project work as the linear dependency, start to start, start to finish, finish to start, and finish to fi nish. With this approach, perceiving the impacts and the variation of res ources and the change of processes is hardly possible. The system dynamics model is used to detail the continuous flow analysis of work process from initial state to the final. This, in tu rns, predicts the effects of performance, scope, and human factors, including resources and processes with project sequence. The system dynamics modeling technique can in corporate the causality links between the variables in a construction system and the ac tivity production process. The model explicitly delineates and simulates the relationships between each variable mathematically. Furthermore, the system dynamics modeling technique allows the construction crew to test the different strategies in a controlled envir onment. One of the most powerful features of system dynamics lies in its analytic capability, which can provide analytic solu tions for complex and nonlinear systems. Table 5-1 shows the notable differenc es between the traditional approach and the system dynamics approach. Why System Dynamics Sterman (1992) and Lee (2006) detailed the an alytic strength of system dynamics in construction projects with respect to understand ing to understand multiple feedback processes in

PAGE 56

56 a project, dealing with soft data, devel oping computer-based model, and managing the complexity and dynamics of large-scale projects. Feedback Process: Feedback processes usually drive th e uncertainty and complexity of construction projects. Understanding the feedba ck process is particul arly important in the strategic decision-making process, because good policy decisions come from exhaustive understanding the system. The ma in idea behind system dynamics is that dynamic and complex behaviors are derived from system structure. System dynamics enables good policy-making and eventually, facilitates the strategic deci sion making process in project management. Aggregation: One of the features of system dynam ics is the ability of the aggregate representation of a project. For example, the stock and flow stru cture, which is a core model structure in system dynamics, can represen t the aggregate behavior. This aggregate representation can contribute to the understanding of the overall system, which in turns enables effective strategies. Once an overall understand ing is developed, a deta iled decision can be supported by adjusting the level of aggregation in system dynamics. Soft Variables: In most construction simulation, the majority of variables are ha rd variables that are available as quantitative metrics and numerica l data. However, most of what we know about the world is descriptive, impressi onistic, and has never been record ed. Thus, soft variables, such as goals, perceptions, and expecta tions, are significant in represen ting the world. Particularly, in strategic construction simulation, soft variables become more important because some policies are derived from these soft variables. That is why system dynamics encourages the use of soft variables in modeling of the strategic decision making process. The wide use of soft variables in system dynamics allows us understand how a policy can be implemented and further, how it can affect construction performance, which in turn, contribute s to determining a good policy.

PAGE 57

57 Computer Modeling: Computer models help overcome ma ny of the limitations of mental models because they are explicit, and their assumptions are open to all for review. The system dynamics model has naturally has many vari ables based on the modelers assumptions. Computer models are able to interrelate these many factors simultaneously and infallibly compute the logical consequences of the assu mptions. The model explicitly delineates and simulates the relationships between each variab le mathematically. Moreover, the computer modeling can be simulated under controlled conditions, allowing analysts to conduct and experiments that are not feasible or ethical in the real system. Large-Scale Projects such as Construction: Large-scale projects such as construction belong to the class of complex dynamics systems. Such systems are extremely complex, consisting of multiple interdependent components, and highly dynamic. Also, these projects involve multiple feedback processes, nonlinear relationships, a nd both hard and soft data. The analytic capability of system dynamics facilitates to re present systems with these characteristics and manage the complexity and dynamics of large-scale projects properly. 5.2 Modeling Process System dynam ics modeling normally follows the five steps; system understanding, conceptualization, formulation of a simulation mo del, validation, and policy design and analysis. Activities of articulating problems to be addressed are conducted in the system understanding step. In this step, key variable s to be considered for a proj ect are introduced, and reference modes are developed. The referen ce mode explains the background of the project and gives the answers to the questions: what is the historical behavior of the key concepts and variables and what might their behavio r be in the future?

PAGE 58

58 Table 5-2. Modeling process System Understanding Problem articulation: What is the pr oblem? What is the purpose of the model? Key variables Reference modes: Specifying the study focus. A set of graphs and other descript data showing the development of the problem over time Conceptualization Dynamic hypothesis: Working theory of explanation of the dynamics characterizing the problem in terms of the underlying feedback and stock and flow structure of the system Model boundary diagrams: The scope of key variables which are endogenous, exogenous, and excluded Formulation of a Simulation Model Causal loop diagrams: The feedback stru cture of systems in th e relation of variables Stock and flow maps: The underlying physical structure of variables accumulating material, money, and information Policy structure diagrams: Causal diagrams showing the information inputs to a particular decision rule Validation Comparison to reference modes Sensitivity Analysis Policy Design and Analysis The creation of entirely new stra tegies, structures, and decision rules What-if analysis Taking into account the purpose of the project and the reference modes, the explanation for model boundary development to scop e the variables to the project, the model for simulation is formulated based on the causal loop diagram and stock and flow maps. That characterizes the state of the system and genera tes the information upon which decisions and actions are based by giving the system inertia and memory (Strerma n 2000). Once the model is completed, validation and analysis steps are needed in accordance wi th the purpose of the project. As one of the strengths of system dynamics approach, the mode l generates several leve ls of scenarios to particularly simulate differen ce decision models. The policy analysis helps to create new strategies, structures, and decision rules for the improvement of project performance.

PAGE 59

59 5.3 Model Boundary A model boundary chart characterizes key variab les in accordance with the scope of the model focused on the modeling purpose. The m odel boundary divides key variables into considerations: which one is included (endogenous), which one is assumed (exogenous), and which one is ignored (excluded). Such a charac terization is important because the clearer the divisions of variable are, th e more successful model is buil t. Endogenous variables are the primary factors of the model to be considered all the tim e of simulation and they can be modified in conformity with the simulation environments. For example, the project duration is one of the considerable key factors for th is research model. Throughout the accelerating schedule in proportion to reducible duration, the duration of th e activity with the redu cible duration will be changed. Finally, depending on the variations of other factors influenced by reduced time, the Table 5-3. Model boundary chart Endogenous Exogenous Excluded Latent lazy time Latency generation and iteration Constraints Activity and project progress Project Duration Activity duration Productivity Schedule compression (pressure) Reliability and stability Evolution External and internal sensitivity Workforce allocation and utilization Work quality Adjusted activity duration Adjusted work scope SC Thoroughness Delay Planned duration Managerial change ratio Buffering size and location Overtime and additional work Request for information rate and time Conflict and dispute Policy choice Change decision time Activity characteristics Dependency Error and change generation and iteration Safety Cash flow Environmental impact Weather and seasonal effect Accident rate Undiscovered change Site condition

PAGE 60

60 total duration that estimated at the planning phase will be changed at the final phase. Exogenous variables are defined and set by us ers and do not change during the simulation. The variables have their important roles for a project such as the planned duration of project, but the values do not change for the analysis of endogenous variables impacts during the simulation. However, if they need to be changed and i nvolved in the major impacts on the simulation, the value can be changed for the purpose of simulation. The excluded variables are cautiously not included in the simulation model because they are beyond the focus of the scope and purpose of si mulation model. The factors such as weather and other environmental impacts ar e hardly anticipated and do not have crucial influence of the project management under schedule compression 5.4 Feedback Processes of Constr uction under Schedule Compression As introduced in Chapter 3, a construction p roject consists of multiple feedback processes, and causal loop diagram is an important tool for representing the feedback structure of systems and showing the dynamics of variables involved in the system. A causal loop diagram comprises variables connected by arrows th at denote the causal influences among the variables, and the important feedback loops are also identified in the diagram (Sterman 2000). A positive (+) sign indicates that an increase (decrease) in one vari able causes a correspondin g increase (decrease) in the dependent variable above what it woul d otherwise have been. On the other hand, a negative (-) sign indicates th at an increase (decrease) Figure 5-1. Concise causal loop diagram of schedule compression and LLT (Lee et al. 2007) Schedule Compression Latent Lazy Time

PAGE 61

61 in the independent variable causes a corresponding decrease (increase) in the dependent variable (Sterman 2000). Fig.5-1 illustrates the relatio nship between schedule compression and LLT. When delay happens and schedule needs to be ac celerated and LLT is det ected through SCT, the removal of LLT is required. Fina lly, LLT is reduced by schedule co mpression. So, this process is represented as the negative relationship. Fig.5-2 delineates the causal loop diagram describing the feedback process possibly existing in general construction. B1, the loop cons isting of delay, schedule pressure, workforce, and production, explains the self -balance loop about the positive impacts of schedule pressure. When delay is detected, schedule should be ac celerated. Appropriate schedule pressure can increase production of project, which is gene rated by increased workforce. And the increased Figure 5-2. Feedback process of general construction

PAGE 62

62 production can be the important role to catch up the delayed schedule. At the same time, R1, the loop of the replacement of production to productivity in B1, indicates the self-reinforcing loop to deteriorate performan ce under schedule acceleration by changes and errors. As in B1 loop, the right amount of schedule pressure can satisfy its purpose to reduce delayed project duration. However, t oo much high schedule pres sure can create too much pressure on workforce, and it results in the decline of produc tivity. For clarity, the concepts of production and producti vity are introduced like this way. Consequently, this cycle can generate another delay, and it requi res another acceleration on schedule. B2 (schedule pressure workforce produc tion delay duration) and R2 (schedule pressure workforce productivity constr uction process duration) are the feedback processes about the rela tionship between schedule pressure and project duration. B2 is the loop to reduce the project duration by compressing schedule process through increasing production. R2 is the negative loop of the ex tension of duration creating addi tional work caused by decreased productivity. Over-accelerated scheduling a nd the inappropriate ex ecution of schedule compression can generate side effects. One of the side effects from schedul e pressure is working out-of sequence. This can create another error and change on plan and reworks are necessary. This delays construction process, and the delay leads alternative actions to meet schedule, the loop R3 of schedule pressure, work out-of-se quence, changes and errors, rework, and delay. Depending on the characteristics of project and the purpose of simulation, causal loop diagram for feedback process can vary. Sche dule pressure could in crease production through more completed work if workers are not fatigue d and do not make mistakes under the situation as in the loop B1. However, schedule pressure may ruin the sequence of work as described in R1. Both two loops increase workforce, but drive op posite results. The structure of causal loop can

PAGE 63

63 be changed depending on the requirement of mode l and the intention of users. For this, the dynamic state of construction caused by feedback processes makes it difficult to create a dynamic model for simulation and to anticipate or measure the constr uction performance. Nevertheless, understanding the construction pro cess with the point of dynamic view based on feedback process enables us to create a unique model for every required situation. The model satisfies the purpose of the project and its mode l, as traditional network-based tools cannot support and the dynamic model makes it convenient to comprehend project with a holistic view. Feedback Process of Construction under Schedule Compression The project management framework analyzed in Chapter 3 is the base of the causal loop diagram of construction proce ss under schedule compression. In this project model, LLT has become one of the most important factors that construction projects inherently have. With the consideration of the delay in the historical data, LLT originat es depending on the accuracy of estimation from the design and planning phase. LLT needs to be detected and managed, which leads to the necessity of schedule compression thoroughness (SCT) (Fig.5-3). SCT, the proc ess to detect LLT in a project, is highly governed by the reliability to identify LLT a nd the information from upstream activity. High reliability increases the correctness of SCT, a nd the correctness identifies more LLT. However, unidentified LLT augments the total LLT of project. Fig.5-4 illustrates the cause-effect relationship of evolution and sensitivity in schedule compression management at the activity level. Th e information flowing from activities based on LLT and latency makes influence on decisions as to whether or not the schedule compression on the activity would be possible and could be executed successfully. If the activity has more tasks that include LLT and require accelerating schedule, the tasks are more sensible to other activities

PAGE 64

64 Figure 5-3. LLT generation loop

PAGE 65

65 Figure 5-4. Internal evolution and sensitivity loop

PAGE 66

66 Figure 5-5. Schedule compression management loop

PAGE 67

67 Figure 5-6. Quality management loop

PAGE 68

68 Figure 5-7. Causal loop diagram of constr uction process under schedule compression

PAGE 69

69 by variation of schedule change. This high sensitivity in activity results in the decrease of construction productivity. In the example of the foundation activity, the task of pouring concrete task highly relies on formwork and pouring concre te task can begin after a portion of completion of formwork. In addition, fractions of the dur ation of formwork are discovered as LLT. Then, some actions for schedule compression are taken to remove LLT, and the change of the upstream task duration, the form work in the example, engenders the fo llowing reprocess of downstream task, the pouring concrete in th e example. Otherwise, the cha nge of the upstream task could generate another error or cha nge of the downstream task. Schedule compression management (SCM) is the process to manage the detected LLT through SCT described in the Chapter 3 and stabili ty decides the possibility of success of SCM. Stability increases the task completeness unde r schedule compression. And, the more tasks are completed in schedule compression, the fewer tasks need schedule compression (Fig.5-5). Meanwhile, schedule pressure is required by the increase of tasks with LLT, which is able to apply schedule compression. However, schedule pressure generates either positive or negative effects. The positive effect results in more tasks completed in schedule compression commonly. But, the negative effect delays projects by th e decrease of productivity from too much pressure on workforce. Furthermore, completed tasks pote ntially have the information for the downstream task, which gives the answer to the question how much work can now be completed based upon how much the work has progressed thus far. Th e increase in the number of tasks makes the completion rate of the downstream tasks increase. Schedule pressure can create not only delays, bu t also other side effects including latency. Latency is the other important concept togeth er with LLT of this model; so, it must be considered separately from other side effect s, and it is highly related to LLT. More LLT

PAGE 70

70 produces more latency in project (Fig.5-6). Thes e side effects require more management through the latency thoroughness process. The credibility of the latency thoroughness process reduces the problem of side effects and th is, in turn, generates more managed side effects and less unmanaged side effects. These processes re cursively influence LLT. Latency thoroughness increases the project quality al though side effects decrease it. The project quality heavily influences the degree of stability-controlling schedule compression process. Finally, through schedule compression management and quality ma nagement processes, LLT can be removed and the removed LLT reduces the project duration as a result. Fig.5-7 shows all feedback process of construction process under schedule compression. 5.5 System Dynamics-Based Project Management Model The generic construction process structure is suited to represent the generic process of construction projects under schedule compression and depicts the dynam ic interactions among variables of activities. In part icular, the process structure provides a model to manage schedule compression in a stabilized work environment a nd to analyze its impacts on the construction performance through three devel opment processes: constraints management process, schedule compression management process, and qua lity management process (Fig.5-8). The causal loop diagrams developed at the prev ious step explain the interdependences and feedback processes of construc tion project. Based on the diagra ms, generic construction process structure is modeled to express quantitative work process consisti ng of stock and flow structure, which represent stored quantities and control qua ntities flowing into a nd out of stocks, or accumulations, respectively. These stocks characteriz e the state of the system and generate the information upon which decisions and actions are based, and stocks are altered by the rates of inflow and outflow (Sterman 2000).

PAGE 71

71 Figure 5-8. Development proces ses in the model structure The model structure provides the methodological approach to contro l construction project under schedule compression through three developm ent processes at wor k, activity, and project levels, and the processes are described with the stock and flow struct ures (Fig.5-13). In the model structure, development works flow into and through six stocks: Work to Do, Work to be Stabilized, Work Awaiting Redesign, Work to be Checked, Work Done not Checked, and Work Completed. Available works from project start-up or the upstream activities are introduced into Work to Do stock through Initial Work Rate for the first time. Before work execution, the introduced works are monitored and checked fo r the existence of constraints and LLT. The discovered constrained works through Constraint Thoroughness activity are carried into the stock of Work to be Stabilized by the number of Constraint Identification Rate and waiting for constraints management process. Constraints management is the process to remove the discovered constraints in works and to stabilize the work environment. Work Stabilization Rate defined by the constraints management process moves the works in Work to be Stabilized stock

PAGE 72

72 into Work to Do stock for execution. The series of this development is constraints management process (Fig.5-10). Stabilized works are scrutinized to detect LLT after constraints management process, and if the works in Work to Do stock do not include LLT, the works are finished through the performance of the work activity, work rate, and then the finished works accumulated in the stock of Work Done not Checked. These works are not inspected as to whether or not they are defective or have not passed through quality ma nagement process. Depending on work quality, fractions of finished works are returned to Work to Do stock through Iterate Work Rate to be corrected, or released to th e downstream activities through Work Completion Rate. This is quality management process. If stabilized works, however potentially have LLT, the works are managed and redesigned to remove LLT and accelerated project schedule. Works with LLT flow into Work Awaiting Redesign stock through Perceived LLT Rate, and schedule compression management redesigns work process and work rate, Reintroduced Work Rate. During schedule compression management, if latency is consider ed at the works in the stock of Work Awaiting Redesign, and then the works are reintroduced to deal with latency through Request to Check Rate and Latency Modified Work Rate. Schedule compression management proce ss includes all of these activities (Fig.5-12). These stock and flow relationships can be desc ribed with the differential equations listed below and equations define how variables of cons truction project work and influence each other. For the complexity and dynamics of construction projects, equations for the simulation of the generic construction process model are represen ted using three dimensional levels, activity, preceding, and succeeding, which are respectively denoted with subscripts j, i, and k, where i, j,

PAGE 73

73 and k{1, 2, n}. These indicate activity itself a nd the relationships of activities with upstream and downstream activities. For clarity, the subscripts j is omitted in the absence of subscripts i and k. (d/dt)(Work to Do) = Initial Work Rate + Work Stabilization Rate + Latency Modified Work Rate + SCT Reprocess Rate + Iterat e Work Rate Work Rate Constraint Identification Rate Perceived LLT Rate Eq. 5-1 (d/dt)(Work to be Stabilized) = Constraint IdentificationWork St abilization Rate Eq. 5-2 (d/dt)(Work Awaiting Redesign) = Perceived LLT Rate Reintroduced Work Rate SCT Reprocess Rate Request to Check Rate Eq. 5-3 (d/dt)(Work to be Checked) = Request to Check Rate Latency Modification Work Rate Eq. 5-4 (d/dt)(Work Done not Checked) = Work Rate + Reintroduced Work Rate Work Completion Rate Iterate Work Rate Eq. 5-5 (d/dt)(Work Completed) = Work Comple tion Rate Eq. 5-6 5.5.1 Basic Work Rate The feedback process in the cau sal loop diagram of c onstruction process, as seen in Fig.5-7, depicts the negative loop consisted of work availa ble, work rate, and work com pleted (described as Task Available, Work Force, and Task Completed respectively in Fi g.5-7), and the causeeffect relationships among these th ree variables decide the work rate. Work rate is based on work available, average work time, and resource constrain. Average work time is the average time required to complete a development work on an activity if all required information, materials, and resources are availa ble and no defects are generated (Ford and Sterman 1998a). Therefore, the work rate mostly relies on work available in average work time, which indicates work produc tivity. Also, work process highly depends on the amount of materials, equipment, and labors. He nce, each activity requires sufficient resources and enough of information. Accordingly, the work ra te of activity is determined as the minimum

PAGE 74

74 Figure 5-9. Initial work rate of resource constraint and work productivity that is the value of work available divided by average work time. In this manner, work to do is introduced by Initial Work Rate, and Initial Work Rate is obtained from Initial Work Available, Initial Work Time, and Initial Resource Constraint. Initial Work Rate = min (Initial Work Availa ble / Initial Work Time, Initial Resource Constraint) Eq. 5-7 5.5.2 Constraints Management Process Before work execution, the works in the stock of Work to Do are considered whether work process is stable or constrained by inappropriate resources or irre gular process thr ough constraint thoroughness. In the similar manner with the basic work rate, Constraint Thoroughness is the lesser of Constraint Thoroughness Restriction and the number of work waiting for constraint thoroughness divided by the average time, which th e process requires to co nstraint thoroughness. Constraint Thoroughness = min (Work to Do / Avg. Constraint Thoroughness Time, Constraint Thoroughness Restriction) Eq. 5-8 Once constraints are under consideration, the works accumulate in Work to be Stabilized stock to remove constraints and stabilize work environment through Constraint Identification Rate. Meantime, constraint thoroughness is not perf ect and all constraint s cannot be discovered.

PAGE 75

75 Figure 5-10. Constraints management process So, Constraint Identification Rate is determined by Discovery Rate of assumed constraints with the fraction of constraints in total work based on Constraint Thoroughness. Therefore, Work Rate of Unconstrained Work is the outflow from Work to Do stock and includes both of stabilized work after constraints management and work with undiscovered constraints. Constraint Identification Rate = Constrai nt Thoroughness Discovery Rate (Total Constraints in Work / Work Scope) Eq. 5-9 Work Rate of Unconstrained Work = Constraint Thoroughness Constraint Identification Rate Eq. 5-10 The discovered constraints in works are ma naged through the constraints management process. The constraints management process wa s explained in the Chapter 4, and Fig.4-3 shows the framework of constraints management. Through the process, Work Stabilization Rate is

PAGE 76

76 defined by constraints management rate and Constraint Subordination. Then, the works are released through Stabilized Work Release Rate relied on Managed Constraint Rate. Work Stabilization Rate = Stabilized Work Re lease Rate Managed C onstraint Rate min (Work to be Stabilized / Avg. Constr aints Management Time, Constraint Subordination) Eq. 5-11 5.5.3 Schedule Compression Management Process The stabilized works released from constrai nts management process accumulate into the stock of Work to Do for Schedule Compression Management (S CM) process. If project managers or workers find any reducible duration in a wo rk or neglected work force (work rate) through Schedule Compression Thoroughness (SCT) process and designate the need of schedule acceleration, the assumed works are carried into the stock of Work Awaiting Redesign through Perceived Latent Lazy Time Rate for rescheduling. Otherwise, works in Work to Do stock proceed with the initially scheduled work process according to Work Rate. This process is determined by the value of LLT Potential. Finally, Perceived LLT Rate is lesser of the SCT process rate and LLT Potential in Work Rate of Unconstrained Work. Perceived LLT Rate = min ((Work to Do / Avg. SCT Time), ((Constraint Thoroughness Constraint Identification Rate) LLT Potential)) Eq. 5-12 Work Rate = min ((Work to Do / Min. Work Time), ((Constraint Thoroughness Constraint Identification Rate) (1 LLT Po tential)), Resource Constraint) Eq. 5-13 Works in the stock of Work Awaiting Redesign are waiting for rescheduling by removed LLT and following schedule compression. The outf low from the stock extremely depends on Schedule Compression Thoroughness, which is the process to monitor and discover LLT. However, SCT cannot detect all of LLT from per ceived LLT. So, the work that possibly does not include LLT returns to Work to Do stock for the reconsideration of SCT or the planned process of work execution. As a result, Schedule Compression Thoroughness Reprocess Rate is determined by incorrectness of SCT upon reprocess time.

PAGE 77

77 SCT Reprocess Rate = (1 SCT) (Work Awaiting Redesign / Min. Reprocess Time) Eq. 5-14 Reintroduce Work Rate is the work rate based on productivity newly generated by schedule compression. This is concerned with the possibility of LLT detection and the stability of rescheduled construction process on schedule compression applications. Reintroduced Work Rate = SCT Stabil ity min ((Work Awaiting Redesign / Avg. Reschedule Time), Schedule Pressure Resource Constraint) Eq. 5-15 However, the adjusted work rate after sche dule compression should alwa ys be larger than the initial work rate. For further information, lets assume there was a project of which the planned duration was 100 days and the total work amount was 100,000 w. (W is used as a hypothetic work unit.) And, 10 days LLT existed in the planned duration. Then, the project should have finished in 90 days. The work enviro nment including work force, material resources, and information of this project had ability to finish the project in 90 days. But, the schedule planned to be done in 100 days. In terms of productivity, 1,000/9 (productivityshould productivityplanned) is wasted as LLT. So, the project should be rescheduled by increasing throughput (a b) in a day work amount based on the productivityshould. The work environment decides arousal, which is interpreted as schedule pressure that determines corresponding productivity. Details of this are shown in the description of the Yerkes-Dodson La w. To remove LLT, the arousal that decides schedule compression should be the optim um level of produc tivity, productivityshould in this case. However, if there is too much schedule pressure on project, the productiv ity in the execution of the project is getting lesser than productivityshould in accordance that the executed arousal is lesser than AO or more. Finally, the productiv ity is lesser than productivityplanned that concludes project delay.

PAGE 78

78 Figure 5-11. Work rates rela tionship by schedule compression For this, Reintroduced Work Rate must be larger than Work Rate. If Reintroduced Work Rate is smaller than Work Rate, the Reintroduced Work Rate would contrarily delay project, which means the uselessness of schedule compre ssion and the need of the reconsideration of SCT. By this reason, the equation 5-15 should be reformulated as: Reintroduced Work Rate = IF THEN ELS E ((SCT Stability min ((Work Awaiting Redesign / Avg. Reschedule Time), Schedule Pressure Resource Constraint)) > Work Rate, (SCT Stability min ((Work Awa iting Redesign / Avg. Reschedule Time), Schedule Pressure Resource Constraint)), 0) Eq. 5-16

PAGE 79

79 Figure 5-12. Schedule compression management process Through Reintroduced Work Rate of schedule compression management process, works with reducible duration or neglected work rate are carried into the stock of Work Done not Checked and waits for quality management process before th e works are released to downstream activities. Under the consideration of schedule compressi on with respect to detecting and removing LLT, once the possible failure of schedule comp ression is identified, the schedule compressed processes should be reconsidered and the possibility of latency shoul d be eliminated. This step is influenced by Schedule Compression Thoroughness and in-Stability of the schedule compression execution to detect LLT. Request to Check Rate = SCT (1 St ability) (Work Await ing Redesign / Min. Reprocess Time) Eq. 5-17 The identified latency is managed and the wo rks after latency modifi cation move back to Work to Do stock for work execution. This step is simila r to the step of constraints management.

PAGE 80

80 Latency Modified Work Rate = min ((Work to be Checked / Avg. Latency Modify Time), Latency Modification Constraint) Eq. 5-18 5.5.4 Quality Management and Work Completed Executed construction works through Work Rate and Reintroduced Work Rate accumulate in the stock of Work Done not Checked to be monitored and inspected. According to the results of quality management, the works in Work Done not Checked stock are either released to the downstream activity or iterated for rework. Quality Assurance (QA) activity is the process to discover defective works. If works are not defective, the works leave Work Done not Checked stock and pass through Work Completion Rate into the stock of Work Completed. If works are found to be defective, the works go back to Work to Do stock through Iterate Work Rate. These works need to be corrected and improved through iterate work activity and return to the initial work stage for reprocesses. In a similar manner to the determination of flow rate, the work available for quality assurance is the number of work s done not checked. Therefore, Quality Assurance rate is the lesser of the work process of quality assurance activity and its resource constraint. Quality Assurance = min ((Work Done not Checked / Ave. QA Time), Quality Assurance Resource Constraint) Eq. 5-19 However, it is impossible to discover all defec tive works by quality assurance activity. Hence, the rate for reprocesses by defects depends on the defects detect rate, Quality Management Thoroughness, and the fraction of defective works, Work to be Iterative Ratio. Accordingly, Iterative Work Rate = Quality Assurance Qual ity Management Thoroughness Work to be Iterated Ratio Eq. 5-20 In the same manner, the Work Completion Rate is determined by non-defective work rate and works not found defects based on Work Release Rate

PAGE 81

81 Figure 5-13. Generic construction pro cess structure under schedule compression

PAGE 82

82 Work Completion Rate = (Quality Assurance (1 Quality Management Thoroughness) (1 Work to be Iterated Ration)) Work Re lease Rate Eq. 5-21 = (Quality Assurance Iterate Work Rate) Work Release Rate Eq. 5-22 5.6 Development of Dynamic Co nstruction Project Model The system dynam ics-based project manageme nt model has been developed through the series of progresses and plays th e important role to understand the association of construction processes under schedule compression. Based on th e system dynamics model, the development of project model enables to identify the interac tion and mechanics of four performance drivers process structure, resources, ta rgets, and scope in a dynamic construction environment. The four performance drivers are strongly relate d to each other and the generic structures for the project, and which is dominant to de cide the project performance. The project performance is implied by the traditional measur es for the construction project: time, quality, and cost. The project model refers them with execu tion, which involves project duration and defects by constraints, project management and cost. The interactions of the primary phase subs ystem are depicted in Fig.514, including project process, resource constr aints, required project goal, and its demands. Each driver has its own components to decide the role in projec t and they influence each other. The generic structure of the system dynamics model and the de pendencies of activities in the project process dominate the project process. For example, in the foundation activity, the task of pouring concrete is totally dependent to the progress of formwork. The stability of the activity according to the generic structure of project and dependency affects the project performance. At the same time, the foundation activity demands concrete and woods for forming and pouring works. These rela tionships are indicated with the solid and dashed lines in Fig.5-14. Demands and requirement s to rely on another dr iver are represented by the solid line. Project performance is mostly influenced by the effectiveness of resource

PAGE 83

83 Figure 5-14. Schema of dynamic project manage ment model (Adapted from Ford et al. 1997) management and can ask the adjustment of res ources. The dashed line explains the return for correctness. Formwork demands woods. So, the woods are supplied for the work as return. Consequently, process structures simulate th e system dynamics model and its dependency. Resource structures simulate the effects and the allocation of workforce, material, and equipments in the development project model. Scope structures model the projected scope of targets for duration, quality, and co st relative to targets and the pr essures on developers due to poor performance (Ford and Sterman 1998a).

PAGE 84

84 CHAPTER 6 APPLICATIONS The previous chapters presented the me t hodology of construction project management under schedule compression and the system dynami cs-based project management model for the purpose of the research has been developed. Base d on the method and the model structures, this chapter scrutinizes the valida tion of the dynamic project mana gement model (DPM) and its applicability as project performance analysis tool by applying the model to a couple of real world case projects. First, the verification focuses on the applicability of DPM with the application of the real construction projects. It explains how to apply the model to real project and examines how close the result of simulation represents to the actual performance. S econdly, the applicability of the model is observed as project performance analysis tool by representing the simulation results in detail. Finally, policy analysis proves the effectiveness of DPM by adapting base case with difference policy scenarios of project performance. These steps help to identify the possibility of the user-defined simulation modeling approach to the planning and control of construction project. For the simulation, Primavera Project Planner (P3), Vensim PLE, and Microsoft Office Excel were used. P3 is a software package for scheduling and tracking pr oject related activities and it worked for the analysis of the pe rformance of the real projects. Vensim is one of the software for system dynamics and used for de veloping, analyzing, and packaging high quality dynamic feedback models. Vensim PLE is the lowest version of the Vensim series. Therefore, there are so many limitations to get sufficient results for the research from Vensim PLE. By this reason, the cooperation with other tools is required. Microsoft Office Excel analyzed the results from the simulation by Vensim PLE.

PAGE 85

85 6.1 Dynamic Project Management Model Application 6.1.1 State Route 25 For the application of DPM to real project, th e construction of highway project in Florida was sim ulated. The project was to widen and rec onstruct State Route 25 (US 27) from north of SR-530 to north of Boggy Marsh Road located in Lake county, Florida (F ig.6-1). The contract was awarded to Ranger Construction Industrie s, Inc. and A+B contract was accepted. The incentive of $5,300 per day was supposed to be paid when the project was finished earlier than original project duration and should not exceed $900,000. The original bid cost was $22 million and the final cost was $22.5 million. This case project serves to verify the validation of DPM by comparison of simulation results with real project perfor mance and the application of DPM for analysis of work progress, the influence of latent lazy time and latency in project and their results. Figure 6-1. SR-25 project location

PAGE 86

86 Table 6-1. Example of input variables for simulation # ID Activity OD2 LLT Stability Sensitivity3 1 1090 Obtain temp drainage material 10 NA NA 9 2 2265 FL power adjustment I 34 4 4 4 3 1010 Install MOT 2 NA NA 9 4 2270 FL power adjustment II 50 5 3 5 5 1140 Install 60 SD trunk line to pond 50 NA NA 9 6 1070 Stage I temp med & RT pav 33 2 NA 6 7 1130 Excavate pond 1 & stockpile 43 4 4 8 8 1160 Clear & grub for stage II 15 3 7 8 9 1155 Excavate pond 2 15 NA NA 9 10 1185 Excavate pond 5 15 6 3 NA .. 81 2110 Install remaining permanent 5 4 3 8 82 2115 Adjust traffic control signal 5 5 2 9 83 2120 Install final roadway signs 5 6 5 8 84 2125 Asphalt friction course 20 5 4 9 85 2130 Final roadway marking 20 5 4 8 86 2135 Remove construction signs 1 NA NA 9 Note: 1. Scale of 0 to 10. Higher value means more LLT included, more stable, and more sensitive to dependency 2. Original Duration 3. External Sensitivity In order to determine the simulation input, a series of interviews with construction managers, superintendents, and scheduler particip ated in the project has been conducted based on their experiences and historical data. 86 major activ ities based on the activities in the critical path was selected for the simulation and. The example of input variables are like as in Table 6-1 and the complete list of input vari ables are shown in Appendix C. Except the major input variables indicated at the above table, DPM has over hundreds variables for simulation to get simulated result close to realistic outcomes. And, some of the variables need to be defi ned as constant value. However, for the simulation of the SR-25 project, the rest of the input variables were given to satisfy the purpose of this research. The comp lete list of the variables and formula for the simulation can be referenced at Appendix B. 6.1.2 Validation The comparison of the simulated output of DPM with the actual project process verifies the applicability of DPM as a pr oject management tool to plan and control construction project

PAGE 87

87 under schedule compression. Also, the result is compared with the Critical Path Method (CPM) planning, which enables to explain the advantag e of the workability of DPM against CPMs. Fig.6-2 illustrates the comparison of the CPM plan, the actual progr ess, and the simulation result of DPM of the SR-25 project at the Percentage Work Complete (PWC) level. The construction of the project started on March 27th, 2006, and the project was planned to finish by January 4th, 2008. The original contract day wa s 650 days based on the analysis of CPM scheduling. The real project, however, finished earlier than 74 days, October 23rd, 2007, the actual construction period of 576 days. And, th e simulated completion date of the project by DPM was November 9th, 2007, 593 days. While CPM plan is behind the actual progress, the simulated output from DPM is almost following the actual progress of the pr oject. As of October 23rd, 2007, the value of Root Mean Square Error (RMSE) between the actual progress and the PWC 100 75 50 25 0 3 3 3 3 3 3 3 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 2 2 2 2 22 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1070140210280350420490560630700 Time (Day)PercentagePWC : CPM 1111PWC : Actual 2222PWC : DPM 3333 Figure 6-2. Percentage of wo rk complete of CPM, actual work, and DPM (SR-25) Oct. 23. 2007 Nov. 9. 2007 Jan. 4. 2008 Mar. 27. 2006 A B

PAGE 88

88 DPM value was estimated at 2.4039% with 2,292 data sets. On the other hand, the RMSE of the actual progress with the CPM plan was estimated at 11.4955%. 6.1.3 Analysis The application of DPM facilitates to analysis project performance in detail and to predict and manage a project based on its results fr om DPM. Fig.6-3 depicts the simulated work progress rate of the SR-25 projec t over the CPM plan. By May 7th, 2007, the project was far behind the plan. Especially, at the initial stag e, the construction proj ect kept delaying with serious decrease rate. After August 22nd of 2006 (day 149), however, th e work progress started to catch up with the original plan. Theref ore, as the turning point of May 7th, 2007 (day 409), the construction progress from DPM surpassed the planned progress, and finally expected to finish 4.95% earlier than the original plan at work progress rate. The identification of the possibility of sc hedule compression is introduced by the work quantity for latent lazy time and latency. In accordance with the schedule compression management process represented in Fig.5-12, the works in the Work Awaiting Redesign stock Work Progress Rate 10 2.5 -5 -12.5 -20 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1070140210280350420490560630700 Time (Day)PercentageWork Progress Rate : Base 1111111111 Figure 6-3. Work progress rate May 7. 2007 Aug. 22. 2006 A B C D Nov. 9. 2007 4.95%

PAGE 89

89 Net Redesigned Work 600,000 450,000 300,000 150,000 0 111 1 1 1 1 1 1 1 1 1 11070140210280350420490560630700 Time (Day)wNet Redesigned Work : DPMBase 111111111Net Work to be Checked 600,000 450,000 300,000 150,000 0 1 1 1 11111111111070140210280350420490560630700 Time (Day)wNet Work to be Checked : DPMBase 11111111 Figure 6-4. Net redesigned work and net work to be checked decide the possibility of work to be redesi gned for schedule compression and also decide the possibility of work to check any latency causing and defects on activities and delay of project. Fig.6-5 demonstrates the cumulative quantity of the work that has possibility for schedule compression and Fig.6-4 illustrates the net redesigned work that could be removed or could be the reason to reduce project durati on by raising work rate according to latent lazy time that the project originally involves and the net work to be checked that could generate delay from changes or errors caused by latency. The simulati on result of DPM predicts that the works for Total Work Awaiting Redesign 100 M 75 M 50 M 25 M 0 1 1 1 1 1 1 1 1 1 1 1 1 11070140210280350420490560630700 Time (Day)wTotal Work Awaiting Redesign : DPMBase 1111111 Figure 6-5. Total work awaiting redesign E F G F E

PAGE 90

90 possible schedule compression converge on around day 410 (E in Fig.6-4), while the works for potential delay by the fail of schedule compression are distributed in the initial and final phases of the project (F and G in Fig.6-4). Further, the concentration of the net redesigned work on E in Fig.6-4 and the net work to be checked on F in Fig.6-4 introduces the rapid increase of cumulative work quantity on E and F in Fig.6-5 (U sing the same alphabet letters in different figures signifies the same causes on the area.) The introduced works by net redesigned work and net work to be checked could reduce the total project duration with elimination of la tent lazy time or expand the duration caused by latency as delineated at the previous chapters in detail. Their effects on productivity are entailed in Fig.6-6, and the total effects of latent lazy time and latenc y on productivity are shown in Fig.6-7. LLT effects on productivity are highly con centrated on H and I in Fig.6-6. According to the net redesigned work, the effects of controll ing LLT increased work performance. It resulted in the boost the total work progr ess and it led the reduction of pr oject duration. A in Fig.6-2 and Fig.6-3 is the consequence of th e LLT effect on productivity, I in Fig.6-6. As the same reason, the LLT effect on H in Fig.6-6 brought the augm ent of C in Fig.6-4 on work progress. Even though LLT effect of H picked more than that of I in Fig.6-6, the reason why the increase of C is LLT Effect on Productivity 10 7.5 5 2.5 0 1 1 1 1 1 1 11 1 1 1 1 1 11070140210280350420490560630700 Time (Day) LLT Effect Ratio : DPMBase 1111111111Latency Effect on Productivity 10 7.5 5 2.5 0 1 1 1 1 1 1 11 1 1 1 1 1 11070140210280350420490560630700 Time (Day) Latency Effect Ratio : DPMBase 111111111 Figure 6-6. LLT and latency effects on productivity H I J K

PAGE 91

91 Total Schedule Compression Effect on Productivity 10 7.5 5 2.5 0 11 1 1 1 1 11 1 11 1 111070140210280350420490560630700 Time (Day) Total Effect : DPMBase 11111111111 Figure 6-7. Total schedule comp ression effect on productivity smaller than the increase of A in Fig.6-3 is that the average LLT effect on A is higher than the average LLT effect on C. Moreover, net redesigned work and LLT effects on productivity started to increase gradually from day 149, and this turn ed into the motive to convert negative work progress to positive one with the ongoing d ecline in latency effects on productivity. LLT effect existed from the initial stage. The total schedule compression effect on productivity, however, did not work at the initial phase of the project, L in Fig.6-7. This is because the redesigned work was curtailed by work to be checked and that consequently eliminated the LLT effect with the absence of net redesigned work even though LLT could have affected the project performance. In addition, the major effect of latency on J in Fig.6-6 is the other reason of no impact of tota l schedule compression on L in Fig.6-7. On the other hand, as the other factor to infl uence project performance, latency effects on productivity drag on project progress under net wo rk to be checked. The impact of latency on productivity exacerbated negative impact on project performance and asked the need of reconsideration of schedule compression. Accord ingly, its concentration on J and K in Fig.6-6 L M

PAGE 92

92 delayed project as seen in B and D of Fig.6-3. The average LLT and latency effects on productivity are respectively 2.0240 and 1.3702. Net work to be checked dose always not indicate the work that causes delay by unsuccessful schedule compression. It contains the work to be reconsidered for schedule allocation. In the case of SR-25 project, the proj ect strolled down in the early phase and caught up with the original plan later. One of the reasons was that many activitie s were planned to begin simultaneously at the early step. When the proj ect launched, however, th e requisite of activity reallocation for better efficiency of work performance was detected. Actually, by moving some activities in the initial phase b ack to later phase, the reduction of project duration was expected and accomplished. This enables to explain the re sult of B in Fig.6-2 and 6-3 and also explain that the net work to be checked of F in Fig.6-4 wa s affected by the latency effect of J in Fig.6-6 and also needed the reconsider ation of rescheduling biased-wor k quantity by the concentrated activities. This elucidation discussed in Chapter 5 as well. The factors discussed so far mutually mani pulate every components of project and that regenerates work rate for proj ect performance. Fig.6-8 states a nd compares the work rates of CPM scheduling and DPM simulation. Latency effects on net work to be checked deteriorated CPM Work Rate 5,000 3,750 2,500 1,250 0 1 1 11 1 1 1 1 1 1 1 1 1 1 1070140210280350420490560630700 Time (Day)w/DayCPM : BaseWorkRate 11111111111DPM Work Rate 5,000 3,750 2,500 1,250 0 1 1 1 1 11 1 1 1 1 1 1 1 11070140210280350420490560630700 Time (Day)w/DayDPM : BaseWorkRate 11111111111 Figure 6-8. CPM vs. DPM work rates N O

PAGE 93

93 the efficiency of planned work rate, N in Fi g.6-8, on the other hand, well-managed LLT on net redesigned work reintroduced a nd increased work rate, O in Fig.68. Therefore, they respectively concluded in the delay and acceleration on projec t progress, B and A in Fig.6-2 and 6-3. While the average work rate of CPM scheduling was 1,500 w/Day, the one of DPM simulation was 1,710 w/Day. W is the work unit for simulation. Reintroduced work rate primarily decided by LLT effect on productivity amplified work rate an d the amplification was the major cause of the reduction on project duration. Consequently, the LLT effects on J in Fig.6-6 cause the augment of total schedule compression effect on M in Fig.6-7, a nd this elicited the major increment of work progress, A in Fig.6-2 and 6-3. The effect of latency on the proj ect, J in Fig.6-6, trailed the initial project progress, B in Fig.6-2 and 6-3. 6.2 Policy Implications The other case project is applied to DPM in order to exam scenario specification in Figure 6-9. SR-25 II project location

PAGE 94

94 different environmental conditions Policy implications design new decision, rules, strategies, and structure that might be tried in the real world and represent how robust policy recommendations are under different scenario an d given uncertainty (Sterman 2000). With this case project, how the time for Schedule Comp ression Thoroughness (SCT) and its reliability affect construction performance are explored. For the objective, another highway project in Florida was simulated, which was the reconstruction of State Route 25 (US 27) from the westbound ramp at SR-50 to County Road 561-A in Lake county of Florida (Fig.6-9) (cal led SR-25 II project in distinction from the previous project). The original bid cost was $25.5 million and the final cost was $27.4 million. The input variables for the simula tion were listed in Appendix C. The project began on February 17th, 2006, and was planned to finish by April 26th, 2008. The original contract day was 800 days based on the analysis of CPM scheduling. However, the project was actually co mpleted by February 16th, 2008, and the actual duration was 730 days. The project duration simulated by DPM was 757 da ys. Fig.6-10 illustrates the PWCs of the CPM PWC 100 75 50 25 0 3 3 3 3 3 3 3 3 3 3 3 3 33 2 2 2 2 2 2 2 2 2 2 2 2 222 1 1 1 1 1 1 1 1 1 1 1 1 1 1 10100200300400500600700800 Time (Day)PercentagePWC : CPM 1111PWC : Actual 2222PWC : DPM 3333 Figure 6-10. Percentage of work complete of CPM, actual work, and DPM (SR-25 II) Schedule CPM Actual DPM Duration 800 days 730 days 757 days

PAGE 95

95 scheduling, the actual work, and the simulation re sult of DPM of SR-25 II project respectively. For 540 days, the DPM progress was behind th e planned progress. As of February 16th in 2008, the estimation of RMSE between the actual progress and the DPM result was 2.5024% with 2,920 data sets. On the other hand, the RMSE of the actual progress with CPM schedule was estimated at 7.5866%. Except the input data, the valu es of many other variables consisting of DPM structure influence simulation results. Most of all, SCT pl ays an important role in the accuracy of the results analyzed by DPM. The time needed for SCT and the reliability of SCT alter the PWC 100 75 50 25 0 4 4 4 4 4 4 4 4 4 44 3 3 3 3 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 10100200300400500600700800 Time (Day)PercentagePWC : 1DaySCT 11PWC : 2DaySCT 22PWC : 1WeekSCT 33PWC : 2WeekSCT 44 PWC 10 7.5 5 2.5 0 4 4 4 4 4 4 4 4 4 4 4 3 3 3 3 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 10102030405060708090100 Time (Day)Percentage"1 Day SCT Time" : DPM 11111111"2 Day SCT Time" : DPM 2222222"1 Week SCT Time" : DPM 3333333"2 Week SCT Time" : DPM 44444444PWC 100 98.75 97.5 96.25 95 4 4 4 4 4 4 4 4 4 4 3 3 3 3 3 3 3 3333 2 2 2 2 2 2 22222 1 1 1 1 1 1 11111700710720730740750760770780790800 Time (Day)Percentage"1 Day SCT Time" : DPM 11111111"2 Day SCT Time" : DPM 2222222"1 Week SCT Time" : DPM 3333333"2 Week SCT Time" : DPM 44444444 (a) Day 0 to 100 (b) Day 700 to 800 Figure 6-11. Percentage of work complete according to SCT time SCT time 1 day 2 days 1 week 2 weeks Duration 757 days 760 days 771 days 797 days

PAGE 96

96 prediction of construction performance. The time a nd reliability of SCT for the base case were 1 day and 100% respectively. Under the condition the simulated duration was 757 days. For the closer results applicable to re al world project, first, the SR-25 II project was simulated under 2 days, 1 week, and 2 weeks of SCT time by DPM. When SCT process took 2 days, the project duration was 760 days. Under SCT process of 1 w eek, the total duration of the project was 771 days, and 797 days of project duration was ge nerated with 2 weeks for SCT process (Fig.6-11). When SCT process needed more than 2 weeks, the simulated duration of DPM was longer than the planned duration by CPM. The base case had 100% of the reliability on SCT. That means that errors or reconsiderations in monitoring and detecting LLT do not exist through SCT process. If SCT process is less reliable, the process will genera te latency resulting in delay of project during schedule compression and iterative cycles caused by errors or reconsiderations will require additional SCT time and errors. For clear c onsideration, the project was simulated under different situations: 0%, 25%, 50% 75%, and 100% of SCT reliabi lity. The results of project PWC 100 75 50 25 0 5 5 5 5 5 5 5 5 4 4 4 4 4 4 4 4 4 3 3 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 10100200300400500600700800 Time (Day)PercentagePWC : 0%SCT 111PWC : 25%SCT 222PWC : 50%SCT 333PWC : 75%SCT 444PWC : 100%SCT 555 Figure 6-12. Percentage of work complete vs. SCT reliability SCT Reliability 0 % 25 % 50 % 75 % 100% Duration 800 days 792 days 781 days 769 days 757 days

PAGE 97

97 Work Rate 4,000 3,000 2,000 1,000 0 5 5 5 5 5 5 5 5 4 4 4 4 4 4 4 4 3 3 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 1 1 1 11 1 1 1 10100200300400500600700800 Time (Day)w/DayWork Rate : 0%SCT 11Work Rate : 25%SCT 22Work Rate : 50%SCT 33Work Rate : 75%SCT 44Work Rate : 100%SCT 55 Figure 6-13. Work rate vs. SCT reliability duration are depicted in Fig.612. The project durations were respectively 800 days, 792 days, 781 days, 769 days, and 757 days on the condition. Th e result with 0% SCT reliability was same as the one of CPM. Which means that CPM sche duling does not consider, even cannot elicit the causes and effects of schedule comp ression on project performance. Fig.6-13 delineates the varied work rates on the different scenarios. As seen in the table, the more reliable SCT is, the higher work rate is. Consequently, highe r work rate decreases project duration. Entirely, the ra te of work rates around day 550 to 600 is high. This is caused by Net Redesigned Work 500,000 375,000 250,000 125,000 0 444 4 4 4 4 4 4 4 333 3 3 3 3 333 2 2 2 2 22 2 2 22 111 1 1 1 1 1 1110100200300400500600700800 Time (Day)w"25%SCT" : WQ 11"50%SCT" : WQ 222"75%SCT" : WQ 33"100%SCT" : WQ 44Net Work to be Checked 500,000 375,000 250,000 125,000 0 4 4 4 44 4 4 4 44 3 3 3 3333 3 3 3 2 2 2 22222 2 2 1 1 1 11111 1 1 10100200300400500600700800 Time (Day)w"25%SCT" : WQ 11"50%SCT" : WQ 222"75%SCT" : WQ 33"100%SCT" : WQ 44 Figure 6-14. Net redesigned work and net work to be checked vs. SCT reliability SCT Reliability 0 % 25 % 50 % 75 % 100% Avg. w/Day 1250 1263 1280 1300 1320

PAGE 98

98 LLT Effect on Productivity 5 3.75 2.5 1.25 0 4 4 4 4 4 4 4 4 4 44 3 3 3 3 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 11 1 1 10100200300400500600700800 Time (Day) "25%SCT" : LLT 11"50%SCT" : LLT 222"75%SCT" : LLT 33"100%SCT" : LLT 44Latency Effect on Productivity 5 3.75 2.5 1.25 0 4 4 4 4 4 4 4 4 4 44 3 3 3 3 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 2 2 1 11 1 1 1 11 1 1 10100200300400500600700800 Time (Day) "25%SCT" : Latency 11"50%SCT" : Latency 22"75%SCT" : Latency 33"100%SCT" : Latency 44 Figure 6-15. Schedule compression effect vs. SCT reliability the LLT effect on productivity within the net redesigned work. Fig.6-14 and Fig.6-15 illustrate the considered work for schedule compressi on and the effect of schedule compression on productivity against differe nt SCT reliabilities. Table 6-2 enumerates the results of durati on, RMSE, and the average of LLT and latency effects under 0% to 100% of SCT reliabilities. As SCT reliability increases, the duration decreases and the accuracy of DPM increases. Th e scarcity of accuracy of DPM reduces the effect of LLT, but augments the impact of latency. Table 6-2. Policy analysis on SCT reliability Reliability Duration (Day) RMSE (%)1 Avg. LLT Effect Avg. Latency Effect 0 % 25 % 50 % 75 % 100 % 800 792 781 769 757 7.5866 6.2164 4.7909 2.6592 2.5024 1.1922 1.2485 1.2662 1.3256 1.2629 1.2785 1.1706 1.0599 Note: 1: 2,920 data sets

PAGE 99

99 CHAPTER 7 CONCLUSIONS Construction is a living organism. Along with the requirem ent of its characteristics, a project is planned and created. Depending on the environment, the growing process of a project varies. The large-scale, uniqueness, and comple xity of a construction project deteriorate the prediction of project progress and its performance, and aggressive project schedules and inappropriate response to delays often invite und esirable consequences on a projects cost and schedule (Nepal et al. 2006). The uncertain ty and variation demand more endeavors of contractors and construction managers as a requi site of schedule compression. Moreover, with the traditional project management methodologies, a lack of understanding of schedule pressure and the perspective consid ering construction as a st atic and linear project make the planning and control of construction under sche dule compression more difficult. As an effort to address these challenge issues, Dynamic Project Management Model (DPM) has been developed in this research. 7.1 Applicability of Dynamic Project Management Model Understanding of Dynamic Construction Process under Schedule Compression Construction process is complex, dynam ic, uns table, and replicate. Multiple feedback processes in a construction projec t contribute to gene rating indirect and unanticipated results during the project execution. A l ack of management mechanism and understanding of schedule compression introduced more effici ent and practical concepts of Latent Lazy Time and Latency for reliable management to accelerate schedule. The development of the frameworks for internal and external management of ac tivities helps understand the association of construction process, latent lazy time, and latency under schedule co mpression and how they affect one another and construction performance in an activity and in a project.

PAGE 100

100 Based on the management frameworks, system dynamics determined the causes and effects of the feedback processes, a nd capturing how development processes affect project performance by explicitly modeling those processes provides significantly improved descriptions of development team mental models, project constr aints, and the drivers of project performance (Ford and Sterman 1998a). To understand the cons truction process with the point of dynamic view based on feedback process enables us to create a unique model for ev ery required situation. The dynamic model satisfies the convenience to co mprehend projects with a holistic view that traditional network-based tools cannot support. Realistic and Proactive Planning System dynam ics offers a project management model that reflects the real experiences of projects, which seldom follow the simple linear route suggested by the logic of the traditional project network. Taking into account dependencies of evolution and sensitivity, production type, and soft data based on feedback processes enable s to augment the reliability of the planning and managing of construction projects. In the case projects for simulation, the project type, the experience level of works, fatigue according to workhours, and the internal and external sensitivities were the critical input variables, and which are also the main input variables of the dynamic project management model and influenc e project performance. By understanding these information, a project manager can use that info rmation to identify appropriate strategies for reducing or removing those dependencies (Bogus et al. 2005), and thus possibly result in the reduction of project duration. The newly introduced concepts, latent lazy time and latency, allow to interpret the quantitative impact of schedule compression on a pr oject. The detrimental impacts of these two factors imply the applicability of the im plementation of schedule compression and the reconsideration of work rate by accelerating schedul e. This interpretation facilitates to predict

PAGE 101

101 project performance and streamlining informa tion flow under schedule compression and the prediction based on the accuracy an d constructability imposes appropriate strategies for better achievement of schedule compression. Through colla borative efforts dire cted at anticipating upcoming results and reducing potential problems in advance, opportunities for the increase in work rate and the decrease in latency can be accomplished. System Dynamics-based Pr oject Management Model With the combination o f Critical Path Method (CPM) scheduling, the DPM based on system dynamics was developed to plan and control construction projects under schedule compression. The user-defined project management model focuses on the ability to dynamically interpret construction processes and practice and analyzes the system structure to understand dynamic behavior for the great contribution to management of a project. The elaboration of the quantitative work for schedule compression and its si de effects, the reintroduced work rate in accordance with schedule acceleration, and the anal ysis of the impacts of LLT and latency on project performance enhances the reliability of prediction and the applicability, and thus devises robust and explicit tactics fo r the best consequences. 7.2 Future Research The dynami c project management model fo r schedule compression that has been developed in the previous steps facilitates to analy ze the relationship of activities and their effects at the activity and project levels a nd to manage project performance under schedule compression. When design and construction beco me more complex and dynamic, concurrent process in the large scale projec t are developed, and projects are globalized, the associated uncertainty and complexity can introduce serious schedule, cost, quality, and safety problems. That increases the factors that must be consider ed and it requires the necessity of more concise

PAGE 102

102 and potent utility to manage the projects. In an effort to address these issues, future researches are directed. 7.2.1 Standardization of Compression Concepts and Methods Project perform ance operated by DPM is pr edicted relied on th e project managers judgment. The input for simulation of DPM such as reliability or stability is different according to the experience and perspective of project ma nagers or experts, which are subjective and ambiguous to make a correct decision to control th e urgent situation on job site, and that creates different outcomes. For this, there is a need to create better objective judgment system into project control. In addition, more systematic and objective efforts to estimate cost and time in the condition of schedule compression must be analyzed than historical data and subjective judgment causing unpredictable results. In 1990, Construction Industry In stitute (CII) at University of Texas Austin published Concepts and Methods of Schedule Compression. As mentioned briefly in Chapter 3, that suggested the 94 concepts and methods for sc hedule compression through broad questionnaire Table 7-1. Effectiveness of methods of sc heduling: schedule value rating (CII 1990) Code Description Value1 Rate2 0-25% CONSTRUCTION 4.01 4.02 4.03 4.04 4.05 4.06 Adaptation to weather conditions Realistic scheduling Repetitive tasks scheduling Schedule crashing Startup-driven scheduling Use of float flexibility 4.2 3.8 4.0 4.1 3.6 3.4 Strong Good Good Good Fair Fair 26-100% CONSTRUCTION 4.01 4.02 4.03 4.04 4.05 4.06 Adaptation to weather conditions Realistic scheduling Repetitive tasks scheduling Schedule crashing Startup-driven scheduling Use of float flexibility 3.9 3.7 3.8 3.8 3.8 3.6 Good Good Good Good Good Fair Note: 1: 5=Strong decrease; 4=Moderate decrease; 3= No effect; 2=Moderate incr ease; 1=Strong increase 2: 4.2-5.0=Strong; 3.7-4.1=Good; 3.3-3.6=Fair; 0.0-3.2=None

PAGE 103

103 Table 7-2. Most effective planned and unpl anned schedule compression methodologies Rank Planned Schedule Compression Unplanned Schedule Compression 1 2 3 15 Staff the project with most efficient crews Detailed project planning Add a second shift Pre-work crew briefings Change to special shifts Staff the project with most efficient crews Detailed project planning Look for short cuts in the process Pre-work crew briefings Avoid interruptions during productive time surveys. In addition to the list and the definition of the schedule compression methods, CII obtained a consensus of opinion from the experienced constructors as to the impact of these 94 concepts and methods on duration and cost duri ng engineering, procurement, and construction. The construction phase was sectioned into Construction 0-25% and Construction 26-100%. Table 7-1 shows the part of the effectiveness of methods of Scheduling rated according to the construction phase. Also, to increase productive time and redu ce the project schedule, Noyce and Hanna delineated planned and unplanned schedule compression based on CIIs schedule compression concepts. This model contains 34 concepts and methods that are determined too most directly apply to the construction phase and the concepts and methods were conducted to explore the impacts of planned and unplanned schedule compression on labor productivity through the questionnaire surveys. The providing informa tion was related to labo r productivity, schedule duration, and project cost impacts. Table 7-2 shows the some parts out of 15 highest ranked results of most effective planned and unpl anned schedule compression methodologies. The methodologies were most effective in maximizi ng labor productivity levels, minimizing costs, and minimizing schedule duration in order (Noyce et al. 1998). By the rapid and magnificent achievement in construction industry over the last decade, however, some of concepts, such as Computer-Aided Design and Drafting or Fast Track Scheduling, have been the norms of strategy. In addition, some of concepts are not suitable to

PAGE 104

104 this research. The premise of this research is to reduce the project duration in the condition of maintenance of time and budget. For example, th e concept of Incentives proposed by CII requires cost increase, which doesnt satisfy the requirements of this research. For these reasons, compression concepts and methods must be reesta blished and reorganized for this research. In the future research, based on the compre ssion concepts of CII, the new and revised compression concepts and methods will be coll ected through diverse literature reviews and practitioner surveys. Then, new compression con cepts and methods will be reorganized and divided into phases by questionnaire surveys: internal management framework, external management framework, design and engineering phases, and construction phases. Through the series of these steps, more practical and rea listic concepts and methods for schedule compression will be determined. 7.2.2 Computerized Dynamic Project Management Model The developed DPM mo del needs a support of dive rse tools to achieve effective results. To schedule and track project related activities, Primavera Project Pla nner (P3) or Microsoft Project was cooperated with Vensim software for the app lication of system dynamics. With the results, Microsoft Office Excel was required as the fi nal analysis of DPM scheduling. Considering complexity and uncertainty of construction pr oject, a series of the processes need a comprehensive system to increase efficien cy and applicability, and thus decrease time consumption in operation of DPM. The computerized DPM intends to integrate se veral existing tools to achieve a flexible application for diverse situation. Particularly, the system dynamics mode l incorporated with network-based and data analysis tools creates user-friendly interf ace so that different kinds and types of input and output are managed in one integrated work stati on. Using well-designed

PAGE 105

105 interface facilitates to translate the simulated results to graphic format for users familiar with CPM. Further, the system increases the accuracy of application in different policies. Figure 7-1. Computerized dynamic project management model Input Activities Duration LLT Stability Sensitivity Output LLT Effect Latency Effect DPM Work Rate

PAGE 106

106 APPENDIX A STRUCTURES OF SYSTEM DYNAMICS MODEL

PAGE 107

107 Figure A-1. Generic construction proce ss structure under schedule compression

PAGE 108

108 Figure A-2. Initial work rate structure

PAGE 109

109 Figure A-3. Work quantity accumulation structure

PAGE 110

110 Figure A-4. Latency modification structure

PAGE 111

111 Figure A-5. Schedule pressure and work rates structure

PAGE 112

112 Figure A-6. Iteration work rate and change and error structure

PAGE 113

113 Figure A-7. Work release rate structure

PAGE 114

114 Figure A-8. Workforce and productivity structure

PAGE 115

115 Figure A-9. Resource co nstraint structure

PAGE 116

116 Figure A-10. Dependency structure

PAGE 117

117 APPENDIX B EQUATIONS OF SYSTEM DYNAMICS MODEL Activity Relationship = Constant (1 for default simulation) Units: Dm nl [0, 2] 1: dependent; 2: interdep endent; 0: independent Actual Productivity = Work Scope / (Work Duration Avg. Workforce) Units: w/(Day*People) Initial productivity for simulation Actual Work Scope = XIDZ (Work Scope, Dependency, 0) Units: w An activity work scope according to its dependency to another activities Avg. Constraint Thoroughness Time = Cons tant (1 for default simulation) Units: Day [0, ?] Avg. Latency Modify Time = Consta nt (1 for default simulation) Units: Day [0, ?] Avg. Quality Assurance Time = Constant (1 for default simulation) Units: Day [0, ?] Avg. Quality Mgmt. Time = Constant (1 for default simulation) Units: Day [0, ?] Avg. Redesign Time = Constant (1 for default simulation) Units: Day [0, ?] Avg. SCT Time = Constant (1 for default simulation) Units: Day [0, ?] Avg. Work Release Time = Constant (1 for default simulation) Units: Day [0, ?] Avg. Work Time = Constant (1 for default simulation) Units: Day [0, ?] Change and Error Detection Rate = C onstant (1 for default simulation) Units: Dmnl [0,1] Change and Error Release Rate = (Change and Error Detection Rate Change and Error Release Rate Amplifier Changes and Errors Hidden Change and Error Quality Mgmt. Thoroughness (Work Release Rate)) / Avg. Quality Mgmt. Time Units: w/Day

PAGE 118

118 Change and Error Release Rate Amplifier = 1e-007 Units: Dmnl [0,1e-007] Changes and Errors = INTEG (Change and Error Release Rate, Constant (10 for default simulation)) Units: w Constraint Identification Rate = Constraint T horoughness Discovery Rate (Total Constraints in Work / Work Scope) Unit: w/Day Constraint Rate = Constant (2 for defau lt simulation) Managed Constraint Rate Units: Dmnl Constraint Subordination = Constr aint Evolution Constraint Classification (1 Other Constraints Subordination) / C onstraint Subordinate Time Units: Dmnl Constraint Thoroughness = MIN ((Work to Do / Avg. Constraint Thoroughness Time), Constraint Thoroughness Restriction) Units: w/Day Constraint Thoroughness Restriction = Workforce Actual Productivity Units: w/Day Dependency = External Depende ncy Productivity Reliability Evolution Fraction Information Sensitivity Constraint Rate Units: Dmnl 1 is normal dependency, more dependent >1 Design Completion Ratio = Constant (1 for default simulation) Units: Dmnl [0, 1] Discovery Rate = Constant (1 for default simulation) Units: Dmnl [0, 1] Downstream Work Scope = Downstream Work Duration Work Unit Converter Units: w Effect Identification Time = Constant (1 for default simulation) Units: Day [0, ?] Effect Rate on Productivity = (((Reliable Produc tivity on Production Type Schedule Pressure Effect on Productivity ) / Fatigue Effect on Productivity) Perceived Productivity) / Avg. Quality Mgmt. Time

PAGE 119

119 Units: w/(Day*People*Day) Even though Work Completed is bigger than PWC, this value doesn't affect Dependency. Only work on less than 1 Evolution Fraction = MAX (1, XI DZ (Work Completed, Perceived Work Consideration, 0)) Standardization Units: Dmnl Even though Work Completed is bigger than PWC, this value doesn't affect Dependency. Only work on less than 1 Expected Work Process = INTEG (Expected Work Rate, Initial Work Process) Units: w Expected Work Rate = MIN (Remaining Work / (Avg. Work Time), Expected Work Rate Constraints) Units: w/Day Expected Work Rate Constraints = Reliabl e Productivity on Production Type Target Workforce Reliability Effect on Work Duration Units: w/Day External Dependency = Lookup for Activity Relati onship (Activity Relationship) Lookup for External Sensitivity (External Sensitivity External Information Exchange) Units: Dmnl External Information Exchange = Co nstant (0 for default simulation) Units: Dmnl [0,1] External Sensitivity = Constant (0 for default simulation) Units: Dmnl [0,1] Fatigue Effect on Productivity = Lookup for Fatigue (Time) Units: Dmnl [0, 10] FINAL TIME = Given Units: Day The final time for the simulation. Fraction of Change and Error = Changes and Errors / Work Completed Units: Dmnl [0,1] Fraction of Work Defects = Work to be Checked / Work to Do Units: Dmnl Hidden Change and Error = MIN ((Total Change and Error Changes and Errors), Work Scope) Units: w

PAGE 120

120 Information Completion Ratio = Design Completi on Ratio Required Work Completion from Upstream Units: Dmnl Information Sensitivity = Lookup for Information Sensitivity (Work Completed Information Completion Ratio) Units: Dmnl Initial Hiring Ratio = Given Units: Dmnl Ratio to hire people per normal pr oductivity. Determine initial workforce Initial Resource Constraint = Cons tant (0 for default simulation) Units: w/Day [0, ?] INITIAL TIME = 0 Units: Day The initial time for the simulation. Initial Work Available = Original Duration Work Unit Converter Units: w Initial Workforce = (Work Scope Initial Hiring Ratio) / (Actual Productivity Work Duration) Units: People Initial Work Process = Constant (0 for default simulation) Units: w [0, ?] Initial Work Rate= MIN ((Initial Work Availabl e / Initial Work Time), Initial Resource Constraint) Units: w/Day Initial Work Time = Given Units: Day Initial Work to Do = Work Duration Work Unit Converter Units: w Iterative Work Rate = Quality Assurance Work to be Iterated Ratio Units: w/Day Latency Modification Constraint = MIN ((Actual Productivity Fr action of Work Defects Motivation to Work Quality Mgmt. Thoroughness ), Initial Work Rate) Units: w/Day

PAGE 121

121 Latency Modified Rate = MIN ((Work to be Check ed / Avg. Latency Modify Time), Latency Modification Constraint) Units: w/Day Latency Modified Rate (for Simulation) = Request to Check Rate + MIN ((Work to be Checked / Avg. Latency Modify Time), Latency Modification Constraint) Units: w/Day LLT Potential = Constant (1 for default simulation) Units: Dmnl [0, 1] Lookup for Activity Relationship ([(0, 0)-(2, 6)], ( 0, 1), (1, 1), (1, 1), (1.3, 1.4), (1.7, 3), (2, 5)) Units: Dmnl Lookup for External Sensitivity ([(0, 0)-(2, 6)], (0, 1), (0.3, 1.4), (0.7, 3), (1, 5)) Units: Dmnl Lookup for Fatigue ([(0, 0) (100, 10)], (0, 1), (100, 1)) Units: Dmnl Lookup for Information Sensitivity ([(0,0)-(200000,10)],(1,1),(70000,1),(100000,1),(200000,1)) Units: Dmnl Lookup for Motivation to Work ([(0, 0) (100, 10)], (0, 1), (100, 1)) Units: Dmnl Lookup for Production Type ([(0, 0) (100, 10)], (0, 1), (100, 1)) Units: Dmnl Lookup for Schedule Pressure ([(0, 0) (100, 10)], (0, 1), (100, 1)) Units: Dmnl Manageable Time Reduction = Work Duration Reliability Units: Day Managed Constraint Rate = Constant (1 for default simulation) Units: Dmnl [0, 1] Min. Reprocess Time = Constant (1 for default simulation) Units: Day [0, ?] Motivation to Work = Lookup for Motivation to Work (Time) Units: Dmnl Original Duration = Given Units: Day

PAGE 122

122 Perceived LLT Rate = MIN ((Work to Do / A vg. SCT Time), (Constraint Thoroughness LLT Potential)) Units: w/Day Perceived Productivity = INTEG (Effect Ra te on Productivity, Ac tual Productivity) Units: w/Day/People with consideration of schedule pr essure, fatigue, and production type Perceived Work Consideration = Work Awaiting Redesign + Work Completed + Work Done not Checked + Work to be Checked + Work to be Stabilized + Work to Do Units: w All stocks in an activity Production Type = Constant (1 for default simulation) Units: Dmnl [0, 1] Determine fast or slow production, closer 1, faster Productivity Reliability = Perceived Productivity / Actu al Productivity Units: Dmnl Quality Assurance = MIN ((Work Done not Chec ked / Ave. QA Time), Quality Assurance Resource Constraint) Units: w/Day Quality Assurance Resource Constraint = Information Completion Ratio Quality Management Thoroughness Workforce Actual Productivity Units: w/Day Quality Management Thoroughness = Constant (1 for default simulation) Units: Dmnl [0, 1] Reintroduced Work Rate = IF THEN ELSE ((SCT Stability MIN ((Work Awaiting Redesign / Avg. Reschedule Time), Schedule Pressure Re source Constraint)) > Work Rate, (SCT Stability MIN ((Work Awaiting Redesign / A vg. Reschedule Time), Schedule Pressure Resource Constraint)), 0) Units: w/Day Release Trigger = IF THEN ELSE (Downstream Work Scope > (Work Done not Checked + Work Completed), 1, 0) Units: Dmnl Reliability = Constant (1 for default simulation) Units: Dmnl [0, 1]

PAGE 123

123 Reliability Effect on Work Duration = Work Dura tion / (Work Duration (Reliability Work Duration)) Units: Dmnl Reliable Productivity on Producti on Type = Actual Productivity (Lookup for Production Type (Production Type)) Units: w/(Day*People) Remaining Work = Target Work Scope Expected Work Process Units: w Request to Check Rate = MIN ((Work Awa iting Redesign / Min. Reprocess Time), (1 Stability) Perceived LLT Rate Schedule Compression Thoroughness) Units: w/Day Required Work Completion from Upstream = (Perceived Work C onsideration of Upstream Work Available of Upstream) / Percei ved Work Consideration of Upstream Units: w Resource Constraint = Informa tion Completion Ratio Resource Limitation Ratio Workforce Actual Productivity Units: w/Day Resource Limitation Ratio = MIN ((Supplied Resour ce Quantity / Target Resource Quantity), 1) Units: Dmnl SAVEPER = TIME STEP Units: Day [0, ?] The frequency with which output is stored. Schedule Compression Thoroughness = Cons tant (1 for default simulation) Units: Dmnl [0, 1] Schedule Pressure Resources Constraint = (1 + XIDZ (MAX (Schedule Pressured Effect on Work Process, 0), Expected Work Process, 0)) Perceived Productivity Units: w/Day Schedule Pressured Effect on Productivit y = Work Duration / Target Duration Units: Dmnl Schedule Pressured Effect on Work Process = Expected Work Process Work Done not Checked Units: Dmnl SCT Reprocess Rate = MIN ((Work Awaiting Re design / Min. Reprocess Time), ((1 Schedule Compression Thoroughness) Perceived LLT Rate)) Units: w/Day

PAGE 124

124 Stability = Constant (1 for default simulation) Units: Dmnl [0, 1] Stable Work Rate for Schedule Compression = MIN ((Work Awaiting Rede sign / Avg. Redesign Time), (Perceived LLT Rate Stability Schedule Compression Thoroughness)) Units: w/Day Standardization = Constant (1 for default simulation) Units: Dmnl [0,1] The level of standardization in th e design product or the design process Supplied Resource Quantity = Constant (1 for default simulation) Units: Dmnl Target Duration = Work Duration Manageable Time Reduction Units: Day Target Hiring Ratio = Constant (5 for default simulation) Units: Dmnl Determine Final Workforce Target Resource Quantity = Constant (1 for default simulation) Units: Dmnl Target Workforce = (Target Work Rate Target Hiring Ratio) / Perceived Productivity Units: People Target Work Rate = Work Available / Target Duration Units: w/Day Target Work Scope = Work Duration ((Work Duration Work Unit Converter) / (Work Duration (Reliability Work Duration))) Units: w TIME STEP = 0.25 Units: Day [0, ?] The time step for the simulation. Time to Adjust Workforce = Constant (1 for default simulation) Units: Day Total Change and Error = Consta nt (10000 for default simulation) Units: w [0, ?] Total Constraints in Work = Cons tant (0 for default simulation) Units: w [0, ?]

PAGE 125

125 Work Available = MAX (0, (Actual Work Scope Perceived Work Consideration)) Units: w High dependency, less work available Work Awaiting Redesign = INTEG (Perceived LLT Rate Request to Check Rate SCT Reprocess Rate Stable Work Rate for Schedule Compression, 0) Units: w Work Completed = INTEG (Work Completion Rate, 0) Units: w Work Completion Rate = Work Release Rate (1 Work to be Iterated Ration)) Units: w/Day Work Done not Checked = INTEG (Work Rate + Re introduced Work Rate Iterate Work Rate Work Completion Rate, 0) Units: w Work Duration = Given Units: Day Workforce = INTEG (Workforce Adjust ment Rate, Initial Workforce) Units: People Workforce Adjustment Rate = (Target Workforc e Workforce) / Time to Adjust Workforce Units: People/Day Work Rate = MIN (MIN ((Work to Do / "Avg. Work Time"), Resource Constraint), Constraint Thoroughness ) Units: w/Day Work Rate of Unconstrained Work = Constraint Thoroughness Constraint Identification Rate Units: w/Day Work Redesigned = INTEG (Stable Work Rate for Schedule Compression, 0) Units: w Work Release Rate = (Release Trigger Qual ity Assurance) / Avg. Work Release Time Units: w/Day Work Scope = Work Duration Work Unit Converter Units: w Work Stabilization Rate = Stabilized Work Re lease Rate Managed C onstraint Rate MIN (Work to be Stabilized / Avg. Constraints Ma nagement Time, Constrai nt Subordination) Units: w/Day

PAGE 126

126 Work to be Checked = INTEG (Request to Check Rate Latency Modified Work Rae, 0) Units: w Work to be Iterated Ratio = Fraction of Change or Error Quality Mgmt. Thoroughness Units: Dmnl [0, 1] Work to be Stabilized = INTEG (Constraint Identification Work Stabilization Rate, 0) Units: w Work to Do = INTEG (Initial Work Rate + Work Stabilization Rate + Latency Modified Rate + SCT Reprocess Rate + Iterate Work Rate Constraint Identification Rate Work Rate Perceived Latent Lazy Time, Initial Work to Do) Units: w Work Unit Converter = 1000 Units: w/Day

PAGE 127

127 APPENDIX C INPUTS FOR SIMULATION Table C-1. Input variables for simulation: SR-25 # ID Activity OD LLT Stability Sensitivity 1 1090 Obtain temp drainage material 10 NA NA 9 2 2265 FL power adjustment I 34 4 4 4 3 1010 Install MOT 2 NA NA 9 4 2270 FL power adjustment II 50 5 3 5 5 1140 Install 60 SD trunk line to pond 50 NA NA 9 6 1070 Stage I temp med & RT pav 33 2 NA 6 7 1130 Excavate pond 1 & stockpile 43 4 4 8 8 1160 Clear & grub for stage II 15 3 7 8 9 1155 Excavate pond 2 15 NA NA 9 10 1185 Excavate pond 5 15 6 3 NA 11 1180 SD trunk line LT 45+30 43 2 3 8 12 1170 SD trunk line LT 20+99 18 3 5 7 13 1630 Grass/sod pond 5 3 7 3 NA 14 1250 SD lat & cross drains LT 45+30 22 3 7 8 15 1190 SD trunk line LT 103+40 28 5 4 7 16 1210 SD trunk line LT 162+00 15 NA NA 9 17 1205 Stabilization LT 45+30 15 4 4 7 18 1255 SD lat & cross drains LT 103+40 15 3 4 8 19 1230 SD trunk line LT 196+00 10 5 6 3 20 1310 Embankment LT 103+40 8 8 2 3 21 1315 Stabilization LT 103+40 12 3 6 9 22 1480 Sidewalk LT 45+30 13 3 7 8 23 1320 Curb & gutter LT 103+40 6 2 6 8 24 1325 Base LT 103+40 10 5 6 9 25 1495 Sidewalk LT 103+40 9 2 4 8 26 1350 Base LT 142+00 6 6 6 8 27 1375 Base LT 162+00 6 7 4 7 28 1515 Gass/sod LT 142+00 2 NA NA 7 29 1400 Base LT 180+00 4 3 6 8 30 1425 Base LT 182+00 4 4 6 7 31 1405 Structural asphalt work LT 4 6 7 5 32 1450 Base LT 209+00 5 5 6 6 33 1455 Structural asphalt LT 209+00 5 4 5 6 34 1655 Pavement making LT 20+99 2 5 4 7 35 1075 TECO people gas adjustments I 54 2 7 5 36 1065 TECO people gas adjustments II 19 4 6 5 37 1660 Install erosion contro l for stage III 5 6 3 6 38 1670 Clear & grub MED 20+90 10 NA NA 7 39 1675 SD cross drains & box culvert 36 5 6 6 40 1680 Median temp pavement 35 4 7 7

PAGE 128

128 Table C-1. Continued # ID Activity OD LLT Stability Sensitivity 41 1685 Install MOT for stage IV 5 NA NA 8 42 1695 Shift traffic for stage IV 1 NA NA NA 43 1700 Clear & grub RT 20+90 25 7 8 6 44 1705 SD Cross drains RT 20+90 3 NA NA 6 45 1710 SD Cross drains RT 39+65 5 3 4 7 46 1730 Embankment RT 20+99 5 2 6 7 47 1735 Embankment RT 39+65 10 5 7 6 48 1740 Embankment RT 80+00 12 4 7 5 49 1760 Stabilization RT 39+65 12 5 7 5 50 2275 FL power adjustments RT 59 5 8 6 51 1765 Stabilization RT 80+00 14 3 8 6 52 1810 Base RT 39+65 12 5 7 5 53 1770 Stabilization RT 125+50 14 2 8 7 54 1790 Curb & gutter RT 80+00 5 6 6 6 55 1815 Base RT 80+00 14 3 7 7 56 1775 Stabilization RT 170+80 14 4 7 6 57 1795 Curb & gutter RT 125+50 5 6 6 7 58 1820 Base RT 125+50 14 3 6 5 59 1840 Structural asphalt Work RT 14 3 7 6 60 1800 Curb & gutter RT 170+80 5 6 6 8 61 2245 Lake groves utilities adjustment 7 6 7 8 62 1825 Base RT 170+80 14 3 8 7 63 1845 Structural asphalt work RT 14 4 6 7 64 1850 Structural asphalt work RT 14 3 7 6 65 1930 Temp pavement markings for stg. V 8 4 5 7 66 1935 Install MOT for stage V 5 6 8 7 67 1945 Shift traffic for stage V 1 NA NA NA 68 1950 Clear & grub MED 20+99 1 NA NA NA 69 2150 Stabilization MED 20+99 3 NA NA 8 70 2155 Stabilization MED 39+65 5 3 4 8 71 2050 Embankment MED 20+99 1 NA NA NA 72 2160 Stabilization MED 80+00 5 3 3 8 73 2165 Stabilization MED 125+50 5 3 4 7 74 2060 Embankment MED 80+00 2 NA NA 6 75 2170 Stabilization MED 170+80 6 2 3 7 76 2065 Embankment MED 125+50 2 NA NA 6 77 1995 Curb & gutter MED 170+80 4 4 2 7 78 2020 Base MED 170+80 4 3 NA 7 79 2045 Structural asphalt MED 170+80 3 NA NA 8 80 2105 Shift traffic to final lane 1 NA NA NA 81 2110 Install remaining permanent 5 4 3 8 82 2115 Adjust traffic control signal 5 5 2 7 83 2120 Install final roadway signs 5 6 5 8

PAGE 129

129 Table C-1. Continued # ID Activity OD LLT Stability Sensitivity 84 2125 Asphalt friction course 20 5 4 9 85 2130 Final roadway marking 20 5 4 8 86 2135 Remove construction signs 1 NA NA 9

PAGE 130

130 Table C-2. Input variables for simulation: SR-25 II # ID Activity OD LLT Stability Sensitivity 1 1030 Clear vegetation 25 NA NA 9 2 1060 Pond excavation 3 83 7 3 7 3 1040 Pond excavation 1 53 6 4 8 4 1070 Final dress and sod pond 1 37 7 3 8 5 1050 Pond excavation 2 25 6 3 8 6 1080 Final dress and sod pond 2 12 5 6 7 7 1090 Final dress and sod pond 3 50 NA NA 3 8 1380 Grade for special 1 15 5 3 7 9 1390 SP Det 1 opt base Gr 09 37 NA NA 7 10 1370 Overbuild asphalt 12 7 3 9 11 1400 Se Det 1 asphalt 11 6 3 9 12 1430 Mill, ARMI, & resurface 18 6 3 8 13 1210 Drainage 18 6 4 8 14 1550 RDWY exc & emb LT 31 5 3 6 15 1760 Final dress & sod LT RDWY 43 4 4 7 16 1890 Ditch pavement 25 6 4 7 17 1720 Stab subgrade 31 3 4 7 18 1730 Curb & gutter 18 4 3 8 19 1750 Sidewalk 37 5 3 6 20 1740 Opt base Gr 09 43 NA NA 7 21 1555 Interconnect cable 37 4 2 8 22 1770 Asphalt 12 4 5 7 23 2600 Median storm drainage 12 3 4 7 24 2620 RDWY embankment 37 6 3 8 25 2630 Stabilization 37 5 4 5 26 2650 Base Gr 09 50 4 4 7 27 2640 Curb & gutter 22 5 5 6 28 2670 Final dress median & sod 31 4 6 7 29 2660 Asphalt 26 4 5 6 30 2680 Shift NB traffic to median 1 NA NA 9 31 2690 Demo O/S shoulder 25 6 5 8 32 2770 Remove guardrail 25 3 4 7 33 2710 RDWY embankment 50 4 3 7 34 2780 Remove curb & gutter 25 5 4 6 35 2730 Sidewalk 31 5 5 7 36 2720 Curb & gutter 18 3 5 6 37 2740 Complete pipe 31 NA NA 6 38 2750 Final dress & sod 31 4 5 7 39 2725 Patch next to curb 12 5 5 8 40 3240 Remove guardrail 12 4 4 8 41 3220 Demo shoulder 18 4 5 8 42 3250 Curb & gutter 25 6 5 9 43 3230 RDWY embankment 31 5 4 8

PAGE 131

131 Table C-2. Continued # ID Activity OD LLT Stability Sensitivity 44 3270 Sidewalk 31 6 3 8 45 3280 Flowable fill next to curb 18 NA NA 9 46 3260 Final dress & sod 37 4 2 9 47 5100 FC-6 30 3 6 8 48 5110 Cure period 12 4 5 9 49 5120 Final striping 12 6 5 9 50 5910 Punch list 10 6 5 9

PAGE 132

132 LIST OF REFERENCES Abudayyes, O., Dibert-DeYoung, A., and jaselski s, E. (2004). Analysis of Trends in Construction Research: 1985-2002. J. Constr. Engr. Mgmt. 130(3), 433-439. Akinci, B., Fischer, M., and Zabelle, T. ( 1998). Proactive Approach for Reducing Non-Value Adding Activities due to Time-space Conflicts. Intl. Group for Lean Construction, Guaruja, Brazil. Assaf, S.A., Al-Khalil, M., and Al-Hazmi, M. (1995). Causes of Delay in Large Building Construction Proejcts. J. Mgmt. Engr., 11(2), 45-50. Ballard, G., and Howell, G. (1994a). Implementi ng Lean Construction: Stabilizing Work Flow. Proc. 2nd Ann. Conf. Intl. Group for Lean Construction, Santiago, Chile. Ballard, G., and Howell, G. (1994b). Implemen ting Lean Construction: Improving Downstream Performance. Proc. 2nd Ann. Conf. Intl. Group for Lean Construction, Santiago, Chile. Ballard, G., and Howell, G. ( 1997). Toward Construction JIT. Lean Construction; 1995 ARCOM Conf., A.A. Balkema, Rotterdam, The Netherlands. Ballad, G., and Howell, G. (1998a). Wha t kind of Production is Construction? Proc. 6th Annual Conf. Intl. Group for Lean Construction, Sao Paulo Brazil, Aug. 13-15. Ballad, G., and Howell, G. (1998b). Shieldi ng Production: Essential Step in Production Control. J. Constr. Eng. Mgmt. 124(1), 11-17. Ballad, G. (2000). The Last Planner System of Production Control. Ph.D. Dissertation, Univ. of Brimingham. Beamon, B.M. (1998). Supply Chain Design and Analysis: Models and Methods. Int. J. Production Economics. 55. 281-294. Bertelsen, S., and Koskela, L. (2004). Cons truction beyond Lean: A New Understanding of Construction Management. 12th Annual Conf. Intl. Group for Lean Construction, Elsinore, Denmark. Bogus, S.M., Molenaar, K.R., and Diekmann, J.E. (2005). Concurrent Engineering Approach to Reducing Design Delivery Time. J. Constr. Eng. Mgmt. 131(11), 1179-1185. Browning, T.R. (1999). Sources of Schedule Risk in Complex System Development. Syst. Engr., 3, 129-142. Bubshait, A.A., and Cunningham, M.J. (1998). C omparison of Delay Analysis Methodologies. J. Constr. Engr. Mgmt., 124(4), 315-322. Chan, W.T., and Hu, H. (2002). Constraint Programming Approach to Precast Production Scheduling. J. Constr. Engr. Mgmt., 128(6), 513-521.

PAGE 133

133 Chang, C., Hanna, A., Lackney, J., and Sullivan, K. (2005). Quantifying the Impact of Schedule Compression on Construction Labor Productivity. Construction Research Congress, Apr. 5-7, San Diego, CA. Choo, H.J., Hammond, J., Tommelein, I.D., Austi n, S.A., and Ballard, G. (2004). DePlan: A Tool for Integrated Design Management. Automation in Constr., 13, 313-326. Choo, H.J., Tommelein, I.D., Ballard, G., and Za belle, T.R. (1999). WorkPlan: ConstraintBased Database for Work Package Scheduling. J. Constr. Eng. Mgmt. 125(3), 151-160. Chua, D.K.H., and Shen, L.J. (2003). Constraint-Based Planning with Integrated Production Scheduler over Internet. J. Constr. Eng. Mgmt. 129(3), 293-301. Chua, D.K.H., and Shen, L.J. (2005). Key Cons traints Analysis with Integrated Production Scheduler. J. Constr. Eng. Mgmt. 131(7), 753-764. Construction Industry Institute (CII). (1990). Concepts and Methods of Schedule Compression. Construction Industry Institu te Publication No. 6-7, The University of Texas at Austin, Austin, TX. Dawood, N. (1998). Estimating Project and Activ ity Duration: A Risk Management Approach using Network Analysis. Constr. Mgmt. Engr., 16, 41-48. East, W.E., and Kim, S. (1993). Standa rdizing Scheduling Data Exchange. J. Constr. Engr. Mgmt., 119(2), 215-225. Eden C., Williams, T., Ackermann, F., and Howic k, S. (2000). The Role of Feedback Dynamics in Disruption and Delay on the Nature of Disruption and Delay in Major Projects. J. Oper. Resear. Socit., 51(3), 291-300. El-Diraby, T.E. (2006). Web-Services Environm ent for Collaborative Management of Product Life-Cycle Costs. J. Constr. Engr. Mgmt., 132(3), 300-313. El-Rayes, K. (2001). Object-oriented Mode l for Repetitive Construction Schedule. J. Constr. Engr. Mgmt., 127(3), 199-205. Eppinger, A.D., Whitney, D.E., Smith, R.P., a nd Gebala, D.A. (1994). A Model-based Method for Organizing Tasks in Product Development. Res. Engr. Design, 6, 1-13. Fiallo, M., and Revelo, V. (2002) Applying the Last Planner C ontrol Sytem to a Construction Project: A Cast Study in Quito, Ecuador. Proc. 2nd Ann. Conf. Intl. Group for Lean Construction, Gramado, Brazil. Ford, D. (2002). Achieving Multiple Project Objectives through Contingency Management. J. Constr. Eng. Mgmt., 128(1), 30-39.

PAGE 134

134 Ford, D., Anderson, S.D., Damron, A.J., Casas, R.L., Gokmen, N., and Kuennen, S.T. (2004). Managing Contractibility Reviews to Reduce Highway Project Duration. J. Constr. Eng. Mgmt., 130(1), 33-42. Ford, D., and Sterman, J. (1998a). Dynamic modeling of product development processes, Syst. Dynamics Rev., 14(1), 31-68. Ford, D., and Sterman, J. (1998b). Expert Knowledge Elicitation to Improve Formal and Mental Models. Syst. Dynamics Rev., 14(4), 309-340. Ford, D., and Sterman, J. (1999). Overcoming the 90% Syndrome: Iteration Management in Concurrent Development Projects. Sloan School of Mangement, MIT. Goldratt, E.M. (1988). Computerized Shop Floor Scheduling. Int. J. Prod. Res., 26(3), 443455. Goldratt, E.M. (1990a). Theory of Constraints, North River Press, New York, NY. Goldratt, E.M. (1990b). What is This Thing Called Theory of Constraints and How Should It Be Implemented?, North River Press, New York, NY. Goodnetllake, J.S., and Garnduff, T.W. (2002) An Integrity Constraint Management Framework in Engineering Design. Computers in Industry. 48. 29-44. Gurley, D., and McManus B., (1998). Practic al Knowledge Builds Projects: Case for Independent Construction Information Management. Intl. Group for Lean Construction, Guaruja, Brazil. Hastak, M., Gokhale, S., Goyani, K. Hong, T., a nd Safi, B. (2007). Project Managers Decision Aid for a Radical Project Cycle Reduction. J. Constr. Engr. Mgmt., 133(6), 437-446. Hegazy, T. (2006). Computerized System for Efficient Delivery of Infrastructure Maintenance/Repair Programs. J. Constr. Engr. Mgmt., 132(1), 26-34. Hegazy, T., and Petzold, K. (2003). Genetic Optimization for Dynamic Project Control. J. Constr. Engr. Mgmt., 129(4), 396-404. Homer, J.B. (1993). A System Dynamics Model of National Cocaine Prevalence. System Dynamics Review, 9(1), 49-78. Horman, M.J., Orposz, M.P., and Riley, D.R. (2006). Sequence Planning for Electrical Construction. J. Constr. Engr. Mgmt., 132(4), 363-372. Howell, G., and Ballard, G. (1994a). "Impleme nting Lean Construction: Reducing Inflow Variation." Proc., 2nd Annual Conf. of the Int. Group for Lean Construction, Santiago, Chile.

PAGE 135

135 Howell, G., and Ballard, G. (1994b). "Lean Production Theory: Moving beyond Can-Do." Proc., 2nd Annual Conf. of the Int. Group for Lean Construction, Santiago, Chile. Howell, G., and Ballad, G. (1998). Implemen ting Lean Construction: Understanding and Action. Proc. 6th Annual Conf. Intl. Group for Lean Construction, Sao Paulo Brazil, Aug. 13-15. Howell, G., and Ballad, G. (1999). "What is Lean Construction?" Proc., 7th Annual Conf., International Group for Lean Construction, Univ. of California, Berkeley, Calif., 1. Howell, G., and Koskela, L. Performing Project Management: The Role of Lean Construction. Hwang, S., and Lie, L. (2005). Proactive Projec t Control using Producti vity Data and Time Series Analysis. Comp. in Civil Engr., The 2005 ASCE International Conference on Computing in Civil Engineering, Lucio Soibelman, July 12, 2005, Cancun, Mexico. Hyari, K., and El-Rayes, K. (2006). Optim al Planning and Scheduling for Repetitive Construction Projects. J. Mgmt. Engr., 22(1), 11-19. Joglekar, N.R., Yassine, A.A, Eppinger, S.D., and Whitney, D.E. (2001). Performance of Coupled Product Development Ac tivities with a Deadline. Mgmt. Sci., 47(12), 16051620. Kandil, A., and El-Rayes, K. (2005). Paral lel Computing Framework for Omptomizaing Construction Planning in Large-Scale Projects. J. Compt. Civil Engr., 19(3), 304-312. Kandil, A., and El-Rayes, K. (2006a). Paralle l Genetic Algorithms for Optimizing Resource Utilization in Large-Scale Construction Projects. J. Constr. Engr. Mgmt., 132(5), 491498. Kandil, A., and El-Rayes, K. (2006b). MACRO S: Multiobjective Automated Construction Resource Optimization System J. Mgmt. Engr., 22(3), 126-134. Koo, B., Fishcer, M., and Kunz, J. (2007). A Fo rmal Identification and Re-sequencing Process for Developing Sequencing Alternatives in CPM Schedules. Automation in Constr., 17, 75-89. Koskela, L. (1992). Application of the New Production Philosophy to Construction. Technical Report No. 72, CIFE, Stanford University, Stanford, CA. Koskela, L. (1999). We Need a Theory of Construction. Berkeley-Stanford CE&M Workshop: Defining a Research Agenda for AEC Process/Product Development in 2000 and Beyond. Stanford, Aug. 26-28. Berkeley, Uni v. of California, Stanford Univ. Koskela, L. (1996). Towards the theory of (lean) construction. Proc. 4th Annual Conf. Intl. Group for Lean Construction,. Birmingham, UK.

PAGE 136

136 Krishnan, V., Eppinger, S.D., and Whitney, D. E. (1997). A Model-Based Framework to Overlap Product Development Activities. Management Science. 43(4), 437-451. Kwak, S. (1995). Policy Analysis of Hanfor d Tank Farm Operations with System Dynamics Approach. Ph.D. Dissertation, Dept. of Nu clear Engr., MIT, Cambridge, MA. Lapinski, A.P., Horman, M.J., and Riley, D.R. (2006). Lean Processes fo r Sustainable Project Delivery. J. Constr. Engr. Mgmt., 132(10), 1083-1091. Lee, D., and Shi, J.J. (2006). Construction Business Automation System. J. Constr. Engr. Mgmt., 132(1), 88-96. Lee, H., Ryu, H., Yu, J., and Kim, J. ( 2005). Method for Calculating Schedule Delay Considering Lost Productivity. J. Constr. Eng. Mgmt., 131(11), 1147-1154. Lee, J., Ellis, R., and Pyeon, J. (2007). Proposed Methodology for Dynamic Schedule Compression. Conference on Automation Science and Engineering 2007 (CASE 2007), IEEE, Phoenix, AZ. Lee, S. (2006). Dynamic Planning and Control Met hodology: Understanding and Managing Iterative Error and Change Cycl es in Large-scale Concurre nt Design and Construction Projects. Ph.D. Dissertation, MIT, Cambridge, MA. Lee, S., and Pena-Mora, F. (2007). Understa nding and Managing Iterative Error and Change Cycles in Construction. System Dynamics Review, 23(1), 35-60. Lee, S., Pena-Mora, F., and Park, M. (2005). Q uality and Change Management Model for Large Scale Concurrent Design a nd Construction Projects. J. Constr. Eng. Mgmt. 131(8), 890902. Lee, S., Pena-Mora, F., and Park, M. (2006a). Reliability and Stabil ity Buffering Approach: Focusing on the Issues of Errors and Changes in Concurrent De sign and Construction Project. J. Constr. Eng. Mgmt. 132(5), 452-464. Lee, S., Pena-Mora, F., and Park, M. (2006b) Dynamic Planning and Control Methodology for Strategic and Operational Cons truction Project Management. Automation in Constr., 15, 84-87. Lee, S., Pena-Mora, F., and Park, M. (2006c). Web-enabled System Dynamics Model for Error and Change Management on Concurrent Design and Construction Projects. J. Compt. Civil Engr.. 20(4), 290-300. Liu, S., and Wang, C. (2007). Optimization M odel for Resource Assignment Problems of Linear Construction Projects. Automation in Construction. 16. 460-473. Lottaz, C., Clement, D.E., Faltings, B.V., and Smith, I.F.C. (1999). Constraint-based Support for Collaboration in Design and Construction. J.Compt. Civil Engr., 13(1), 23-35.

PAGE 137

137 Lu, M., and Lam, H. (2008). Critical Path Scheduling under Resource Calendar Constraints. J. of Const. Engr. and Mgmt., 134(1), 25-31. Luebbe, R., and Finch, B. (1992). Theroy of Constraints and Lin ear Programming: a Comparison. Int. J. Prod. Res. 30(6), 1471-1478. Lyneis, J.M., Gooper, K., and Els, S.A. (2001). Strategic Management of Complex Projects: A Case Study using System Dynamics. System Dynamics Review, 17(3), 237-260. Lyneis, J.M., and Ford, D.N. (2007). System Dynamics Applied to Project Management: A Survey, Assessment, and Directions for Future Research. System Dynamics Review, 23(2/3), 157-189. Martin, A., and Formoso, C. (1998). Evaluatio n Building Systems based on Production Process Management and Lean Construction Concepts. Intl. Group for Lean Construction, Gramado, Brazil. Miller, J.B., Garvin, M.J., Ibbs, C.W., and Mahoney, S.E. (2000). Toward a New Paradigm: Simultaneous Use of Multiple Project Delivery Methods. J. Mgmt. Engr., 16(3), 58-67. Moselhi, O. (1993). Schedule Compressi on using the Direct Stiffness Method. Can. J. Civ. Eng. 20, 65-72. Moselhi, O., and Hassanein, A. (2003). Op timized Scheduling of Linear Projects. J. Constr. Eng. Mgmt., 129(6), 664-673. Motawa, I.J., Anumba, C.J., Lee, S., and Pena-M ora, F. (2007). An Integrated System for Change Management in Construction. Auto. Constr., 16, 368-377. Nasir, D., McCabe, B., and Hartono, L. (2003). Evaluating Risk in Construction-Schedule Model (ERIC-S): Construction Schedule Risk Model. J. Constr. Engr. Mgmt., 129(5), 518-527. Nepal, M., Park, M., and Son, B. (2006). Eff ects of Schedule Pressure on Construction Performance. J. Constr. Eng. Mgmt. 132(2), 182-188. Nitithamyong, P., and Skibniewski, M.J. (2006). Success/Failure Factors and Performance Measures of Web-Based Construction Proj ect Management Systems: Professionals Viewpoint. J. Constr. Eng. Mgmt., 132(1), 80-87. Noyce, D.A., and Hanna, A.S. (1997). Planned Schedule Compression Concept File for Electrical Contractors. J. Constr. Eng. Mgmt. 123(2), 189-197. Noyce, D.A., and Hanna, A.S. (1998). Planne d and Unplanned Schedule Compression: the Impact on Labor. Construction Management and Economics, 16, 79-90. Ogunlana, S., Lim, J., and Saeed, K. (1998). Desman:A Dynamic Model for Managing Civil Engineering Design Projects. Computers and Structures, 67, 401-419.

PAGE 138

138 Ogunlana, S., Li, H., and Sukhera, F.A. (2003). System Dynamics Approach to Exploring Performance Enhancement in a Construction Organization. J. Const. Engr. and Mgmt., 129(5), 528-536. Park, M. (2001). Dynamic Planning and Cont rol Methodology for La rge-Scale Concurrent Construction Projects. Ph.D. Dissertation, MIT. Park, M. (2005). Model-based Dynamic Res ource Management for Construction Projects. Auto. Constr., 14, 585-598. Park, M., and Pena-Mora, F. (2003). Dynami c Change Management for Construction: Introducing the Change Cycle into Model-based Project Management. Sys. Dyn. Review, 19(3), 213-242. Park, M., and Pena-Mora, F. (2004). Reliab ility Buffeing for Cons truction Projects. J. Constr. Eng. Mgmt., 130(5), 626-637. Pena-Mora, F., and Li, M. (2001). Dyna mic Planning and Control Methodology for Design/Build Fast-Track C onstruction Projects. J. Constr. Eng. Mgmt., 127(1), 1-17. Pena-Mora, F., and Park, M. (2001). Dynamic Planning for Fast-Track Building Construction Projects. J. Constr. Eng. Mgmt., 127(6), 445-456. Pena-Mora, F., and Dwivedi, G.H. (2002). M ultiple Device Collaborative and Real Time Analysis System for Project Mana gement in Civil Engineering. J. Compt. Civil Engr., 16(1), 23-38. Pena-Mora, F., and Tanaka, S. (2002). Inf ormation Technology Planning Framework for Japanese General Contractors. J. Mgmt. Engr., 18(3), 138-149. Pheng, L.S., and Tan, S.K.L. (1998). How Justin-Time Wastages can be Quantified: Case Study of a Private Condominium Project. Constr. Mgmt. Econ., 16, 621-635. Prasad, B. (1996). Concurrent Engineering Fundamenta ls: Integrated Product and Process Organization, Prentice Hall, Upper Saddle River, NJ. Rahman, S. (1998). Theory of C onstraints: A Review of the Philosophy and Its Application. Int. J. Operat. Product. Mgmt. 18(4), 336-355. Rahman, S. (2002). Theory of Constraints Thinking Process Approach to Developing Strategies in Supply Chains. Int. J. of Physical Di stribution and Logistics Mgmt. 32(9/10), 809-828. Ren, Z., Anumba, C.J., and Ugwu, O.O. (2003). Multiagent System for Construction Claims Negotiation. J. Compt. Civil Engr., 17(3), 180-188. Richardson, G.P., and Pugh, A.L. (1981). Introduction to System Dynamics Modeling with DYNAMO, MIT Press.

PAGE 139

139 Richardson, G.P. (1985). Introduction to the System Dynamics Review. System Dynamics Review, 1(1), 1-5. Rodrigues, A., and Bowers, J. (1996a). S ystem Dynamics in Project Management: A Comparative Analysis with Traditional Methods. System Dynamics Review, 12(2), 121139. Rodrigues, A., and Bowers, J. (1996b). The Role of System Dynamics in Project Management. Int. J. of Porject Mgmt., 14(4), 213-220. Royset, J.O., Kiureghian, A.D., and Polak, E. (2006). Optimal Design with Probabilistic Objective and Constraints. J. Engr. Mech., 132(1), 107-118. Sacks, R., Esquenazi, A., and Goldin, M. (2007) LEAPCON: Simulation of Lean Construction of High-Rise Apartment Buildings. J. Constr. Engr. Mgmt., 133(7), 529-539. Scheer, S., Santos, A., Quevedo, J.R.S., Filho, J. M., and Boica, S.M.R. (2005). IT as a Success Factor for a Lean Construction System in a Precast Concrete C onstruction Industry. Constr. Civil Engr., 179, 160. Simatupang, T.M., Wringt, A.C., and Sridharan, R. (2004). Applyi ng the Theory of Constraints to Supply Chain Collaboration. Supply Chain Mgmt. 9(1), 57-70. Sterman, J. (1992). System Dynamics Modeling for Project Manageme nt. Sloan School of Mangement, MIT, http://web.mit.edu/jsterman/www/. Sterman, J. (2000). Business Dynamics: System Thinki ng and Modeling for a Complex World. McGraw-Hill Companies, New York, NY. Steyn, H. (2002). Project Management Applicat ions of the Theory of Constraints beyond Critical Chain Scheduling. Int. J. Project Mgmt., 20, 75-80. Svenson, O, and Maule, A. J. (1993). Time Pressure and Stress in Human Judgment and Decision Making. Plenum, New York. Teicholz, P. (2001). U.S. Construc tion Labor Productivity Trends, 1970-1998 J. Constr. Eng. Mgmt., 126(2), 427-429. Thomas, H.R. (2000). Schedule Accelerati on, Work Flow, and Labor Productivity. J. Constr. Eng. Mgmt., 126(4), 261-267. Thomas, H.R., and Horman, M. (2006). Fundament al Principles of Workforce Management. J. Constr. Eng. Mgmt., 132(1), 97-104. Thomas, H.R., Horman, M., and Lemes de Souza, U. (2004). Symbiotic Crew Relationships and Labor Flow. J. Constr. Eng. Mgmt., 130(6), 908-917.

PAGE 140

140 Thomas, H.R., Horman, M., Lemes de Souza, U., a nd Zavrski, I (2002). R educing Variability to Improve Performance as a L ean Construction Principle. J. Constr. Eng. Mgmt., 129(3), 251-261. Thomas, H.R., Horman, M., Minchin, R., a nd Chen, D. (2003). Improving Labor Flow Reliability for Better Productivity as Lean Construc tion Principle. J. Constr. Eng. Mgmt., 129(3), 251-261. Thomas, H.R., and Naplitan, C.L. (1995). Quan titative Effects of Construction Changes on Labor Productivity. J. Constr. Engr. Mgmt., 121(3), 290-296. Thomas, H.R., and Sakarcan, A.S. (1993). Forec asting Labor Productivity using Factor Model. J. Constr. Engr. Mgmt., 120(1), 228-239. Thomas, H.R., and Zavrski, I. (1999). C onstruction Baseline Produc tivity: Theory and Practice. J. Constr. Eng. Mgmt., 125(5), 295-303. Tommelein, I.D. (1998). Pull-Driven Scheduling for Pipe-Spool Installation: Simulation of a Lean Construction Technique. J. Constr. Engr. Mgmt., 124(4), 279-288. Tommelein, I.D., and Ballard, G. (1997). Looking-ahead Planning: Screening and Pulling. Tech. Rep. No.97-9, Civil and Environmental Engineer ing Dept., Univ. of California, Berkeley, Calif. Tommelein, I.D., and Ballard, G. (1998). Coordinating Specialists. Submitted to J. Const. Engr. Mgmt. Tsi, M., Yang, J., and Lin, C. (2007). Synchroni zation-based Model for Improving on-site Data Collection Performance. Auto. Constr., 16, 323-335. Turek, M. (1995). System Dynamics Analysis of Financial Factors n Nuclear Power Plant Operations. MS Thesis, Dept. of Nuclear Engineering, MIT, Cambridge, MA Vrijhoef, R., and Koskela, L. (2000). The Four Roles of Supply Chain Management in Construction. European J. of Purchasing and Supply Mgmt., 6. 169-178. Wickens, C.D., and Hollands, J.G. (2000). Engineering Psychology and Human Performance. 3rd Ed., Prentice Hall, Upper Saddle River, NJ. Womack, D.E., and Flowers, S. (1999). Improving System Performance: A Case Study in the Application of the Theo ry of Constraints. J. of Healthcare Mgmt., 44(5), 397-407. Womack, J., and Jones, D. (1996). Lean Thinking. Simon and Schuster, New York. Zhang, C., Hammad, A., Zayed, T.M., Wainer, G., and Pang, H. (2007). Cell-based Representation and Analysis of Spatial Resources in Construction Simulation. Automation in Construction, 16, 436-448.

PAGE 141

141 BIOGRAPHICAL SKETCH Jaesung Lee was born in 1974, in Busan, South Korea, and rema ined there until he graduated from Dong-A University in 2001. He had studied Architectural Engineering as an undergraduate student in Busan, South Korea, an d had completed over two years of a research student in the construction engi neering laboratory in Dong-A Univ ersity. In the period as a research student, he worked some projects like as structural safety survey and assessment and research in construction management. After graduation, he immediately was eager to pursue the advanced studies in the United States. He enrolled in the Fu Foundation Sc hool of Engineering and Applied Science at Columbia University in the city of New York and studied the diverse areas from planning and programming, estimating, scheduling, and coordinati on and control to strate gies for construction management through a lot of case studies under professor Michael J. Garvin, a Ph.D. in construction engineering and management from MIT. Especially, his in terests and knowledge included sustainable strategies for infrastructure delivery and management and engineering and construction markets and organization and re searched the strategies for construction organizational evolution. After receiving the masters degree from Columbia University, he transferred to the University of Florida for Ph.D. to satisfy his thirst for knowledge about this in 2003. He had researched several projects as a graduate rese arch assistant under Dr. R. Ralph Ellis and Dr. Edward Minchin. Also, he taught the class of te chnical drawing and visualization as a teaching assistant.